Home 🛠️ Troubleshooting

🛠️ Troubleshooting

Camilo Aponte
By Camilo Aponte
• 67 articles

🧠 Brand Vault

This article covers: - Brand Vault auto-population - brand voice analysis - AI personalization - style guide uploads - storage limitations - known issues and bug workarounds - Knowledge Graph social profile mapping - OTTO activation prerequisites - and the Brand Vault icon in Campaign Hub. Brand Vault acts as the centralized memory and identity layer across Search Atlas AI systems. Because multiple tools depend on Brand Vault data asynchronously, updates may require: - reprocessing - re-indexing - OTTO synchronization - or fresh AI generation sessions before changes appear everywhere. ⚠️ Error 1: Brand Vault AI Auto-Populate is not filling in my brand information What's happening The Brand Vault crawler may struggle with: - JavaScript-heavy websites - bot protection systems - Cloudflare restrictions - incomplete metadata - or inaccessible pages. This is especially common on: - React - Next.js - Vue - and heavily client-rendered websites. Steps to try 1. Confirm your website URL is correct. 2. Open the website in an incognito browser to verify it loads publicly. 3. Make sure the homepage is not behind: - logins - paywalls - captchas - or geo-blocking. 4. If your site is JavaScript-rendered, manually complete: - Business Info - Products/Services - Brand Description - and Voice Instructions. 5. Save all manual changes before retrying Auto-Populate. 6. If some sections are already filled in manually, avoid running Auto-Populate on the full vault as it may overwrite your entries. Instead, only trigger Auto-Populate on sections that are completely empty. 7. Remove temporary bot protection if possible. 8. Ensure metadata exists on the site: - title tags - meta descriptions - structured business information. 9. Retry the crawl after several minutes. 10. Contact support if auto-populate consistently fails across multiple sections. Important Notes - Manual entries improve future AI generation quality immediately. - Auto-Populate works best on publicly crawlable HTML content. ⚠️ Error 2: Brand voice analysis returns a generic result What's happening Brand voice analysis requires meaningful source material to detect: - tone - writing patterns - vocabulary - positioning - and communication style. Without enough source material, the AI defaults to generic voice modeling. Steps to try 1. Go to: More Features → Brand Vault → [your brand] → Vault → Add Sources 2. Upload at least 3–5 high-quality writing samples. 3. Add: - blog posts - landing pages - newsletters - press releases - or existing website content. 4. Avoid uploading extremely short text samples. 5. Use content that reflects your actual brand tone. 6. Remove outdated or low-quality samples. 7. Re-run Analyze Brand Voice after adding sources. 8. Add both long-form and short-form writing examples. 9. Refine Voice Instructions manually after analysis. 10. Generate a small test article to validate tone accuracy. Important Notes - Better source quality directly improves AI tone matching. - Brand voice improves progressively as more content is added. ⚠️ Error 3: Content generated by AI tools does not match my brand voice What's happening Brand Vault settings only apply when: - the correct vault is selected - the content tool supports Brand Vault integration - and the generation session starts after the vault configuration is loaded. Steps to try 1. Confirm the correct Brand Vault is selected in the AI tool. 2. Open the Brand Vault selector before generation. 3. Verify Voice Instructions are fully completed. 4. Add additional writing samples if tone remains generic. 5. Start a completely new generation session after updating the vault. 6. Avoid reusing older drafts created before vault updates. 7. Refine: - tone - audience - brand personality - and writing constraints. 8. Use more specific prompts during generation. 9. Test smaller content generations first. 10. Apply Updated Settings inside existing drafts when available. Important Notes - Existing drafts may still use older voice settings. - Voice consistency improves after multiple generations. ⚠️ Error 4: Style Guide images not uploading What's happening Brand Vault image uploads only support: - JPG - PNG - WebP - and GIF formats. Uploads fail when: - files exceed size limits - unsupported formats are used - or metadata corruption exists. Steps to try 1. Confirm the image format is: - JPG - PNG - WebP - or GIF. 2. Convert HEIC or TIFF files before upload. 3. For image uploads, resize or compress images larger than 10 MB. 4. Remove unusual characters from the filename. 5. Retry the upload in Chrome. 6. Disable browser extensions temporarily. 7. Refresh the Brand Vault page. 8. Upload one image at a time. 9. Test with a smaller image file. 10. Contact support if uploads fail for all supported formats. Important Notes - Large design files frequently exceed upload limits. - PNG and JPG formats are the most stable. Confirmed upload limits - Documents: up to 200 MB per file and up to 10 files in one upload. Supported formats are TXT, TEXT, MD, MARKDOWN, PDF, DOC, DOCX, and XLSX. - Style Guide images: PNG, JPG, GIF, or WEBP; up to 10 MB per image and up to 100 images in one upload. - The live upload interface does not show a total source-count limit. ⚠️ Error 5: Storage limit banner is showing — cannot add more sources What's happening Each Brand Vault has a storage quota based on: - subscription plan - uploaded source files - image usage - and stored brand assets. Steps to try 1. Go to: Brand Vault → [your brand]. ⚠️ Error 6: Social media links are disappearing or not saving What's happening This is a known backend bug related to the social_profiles JSONB mapping. Social media URLs entered into Brand Vault may fail to persist across accounts, affecting multiple agency setups. Steps to try 1. Avoid re-entering the same social media links repeatedly — they will not persist while the bug is active. 2. Contact support so your case can be linked to the active engineering fix. Important Notes - This issue is tracked under engineering ticket SPE-732. - Support will notify affected accounts once the fix ships. ⚠️ Error 7: Brand Vault images or logos appear missing or cannot be clicked What's happening Images, logos, and Style Guide uploads may appear missing or unresponsive on the frontend even though the files are still intact on the backend. Clicking the images may produce no response and no error message. Steps to try 1. Clear your browser cache and cookies. 2. Test the page in an incognito browser window. 3. Log out and log back in. 4. Contact support to request a backend check if the issue persists. Important Notes - Affected images are typically still stored on the backend and are not truly deleted. - Support can confirm whether your files exist server-side and restore frontend visibility. ⚠️ Error 8: Brand Vault preview sections show generic or irrelevant text What's happening Vault → Preview → Product and Vault → Preview → Support sections may display generic, software-company-style text that does not match the actual business. This is a known auto-population bug affecting certain business types. Steps to try 1. Manually overwrite the preview text in the Product and Support sections with content that accurately reflects your business. 2. Save the corrected preview before generating AI content. Important Notes - The underlying defect has been escalated to engineering. - Manual overwrites will persist until the auto-population logic is corrected. ⚠️ Error 9: Cannot create Brand Vault — domain or name already exists What's happening You may see the message "Brand Vault with this domain and name already exists. Please use a different domain or name." even when no duplicate vault exists. This is a known backend bug where the duplicate-check incorrectly blocks creation. Steps to try 1. Contact support to have the false conflict resolved manually. 2. Avoid creating workaround vaults under altered names — the official fix is awaiting release. Important Notes - A permanent fix is awaiting release. - Support can clear the duplicate-check conflict on your account in the meantime. 📌 Known Issues: Knowledge Graph & OTTO Activation Knowledge Graph social profile mapping Social profile URLs in the Knowledge Graph (BBB, LinkedIn, Yelp, GBP) may appear in incorrect fields due to a backend JSONB mapping bug. Verify the field assignments after auto-population and contact support if values are misplaced. OTTO activation prerequisites Brand Vault data integrity issues — including missing images, dropped social profiles, or Knowledge Graph mapping errors — can prevent OTTO from activating correctly. Resolve the underlying Brand Vault defects before retrying OTTO activation. 💡 Tip: Brand Vault icon in Campaign Hub In Campaign Hub, businesses with a linked Brand Vault now display a small Brand Vault icon next to their name. This gives you a quick visual indicator of which businesses are connected to a Brand Vault without having to open each one.

Last updated on Oct 26, 2026

🛠️ How Atlas Brain Routes AI Requests — and What Causes Errors in Atlas Agent

🛠️ How Atlas Brain Routes AI Requests — and What Causes Errors in Atlas Agent Every task you send to Atlas Agent passes through Atlas Brain — the AI routing engine that selects, calls, and returns results from external AI model providers. When Atlas Brain encounters a problem at the routing layer, Atlas Agent displays an error message. While many of these errors are transient and resolve within minutes, some may persist for 24–48 hours during platform incidents — if an error continues for more than 30–60 minutes, contact support rather than waiting. Understanding how this routing works helps you interpret errors faster, choose the right next step, and know when to wait versus when to contact support. 🧠 What Is Atlas Brain? Atlas Brain is the core AI engine that powers Atlas Agent. It acts as an intelligent dispatcher sitting between your request and the external AI models that fulfill it. Every Atlas Agent response you receive has been routed, processed, and returned by Atlas Brain. Without a functioning Atlas Brain connection, Atlas Agent cannot complete any task. Atlas Brain is not a single AI model — it is a routing and orchestration layer that manages connections to multiple external AI model providers simultaneously, selecting the best available model for each request type. 🔀 How Request Routing Works Step by Step ➡️ Step 1 — Task received: You submit a task to Atlas Agent. Atlas Agent captures your input and forwards it to Atlas Brain. ➡️ Step 2 — Model selection: Atlas Brain evaluates the task type and selects the most appropriate external AI model provider for that specific request. ➡️ Step 3 — Provider API call: Atlas Brain sends the request to the selected provider's API endpoint and waits for a response. This is the step where failures occur. ➡️ Step 4 — Response delivered: The provider's output is processed by Atlas Brain and returned to Atlas Agent as your completed result. Under normal conditions this entire flow completes in seconds. An error surfaces when Step 3 fails — Atlas Brain cannot reach the provider, the provider returns an error, or the target model no longer exists at the expected endpoint. ⚠️ The Three Root Causes of Atlas Brain Routing Failures 🔧 Provider-side maintenance: External AI providers occasionally take their APIs offline for planned or emergency maintenance windows. Atlas Brain detects the unavailability immediately and returns an error rather than leaving your request in a permanent pending state. These outages are typically brief — most resolve within two to five minutes — and Atlas Brain resumes automatically once the provider is back online. 🚦 Rate limits: AI providers enforce usage ceilings on their APIs. If Atlas Brain reaches a provider's rate limit during a period of high demand, requests temporarily fail until the rate window resets. Like maintenance outages, this is a transient condition that self-resolves without any action on your part. ❌ Provider infrastructure failures: Unexpected faults — regional outages, degraded APIs, or transient network issues on the provider's infrastructure — can interrupt the connection between Atlas Brain and the provider. These are outside Search Atlas's direct control and resolve when the provider restores service. They typically appear as intermittent failures rather than a complete outage. 🔴 Model removal or renaming (the most impactful cause): When an AI provider retires, renames, or restructures a model, Atlas Brain's routing configuration still points to the old endpoint — which no longer exists. This causes every request to fail immediately and consistently, not just occasional requests, because the model is permanently gone. This failure type does not self-resolve. It requires the Search Atlas engineering team to update the routing configuration and deploy a replacement model. 🔁 What the Failure Pattern Tells You ⏱️ Intermittent pattern — the error appears on some requests but not all, and clears within two to ten minutes. This almost always indicates a transient provider issue: maintenance, a rate limit reset, or a brief infrastructure fault. Waiting and retrying is the correct response. Self-service troubleshooting steps resolve these cases. 🔴 Persistent pattern — the error fires on every single request from the start and does not recover after ten or more minutes. This strongly indicates a routing misconfiguration, most commonly a retired or renamed model on the provider side. No browser-side troubleshooting step will resolve this. The fix must be applied on the routing layer by the engineering team. Identifying which pattern you are experiencing is the single most useful piece of information you can provide when contacting the Search Atlas support team. 👷 How the Search Atlas Engineering Team Monitors and Responds 🔍 For transient provider failures: Atlas Brain's monitoring detects elevated error rates automatically. The engineering team watches provider status pages and internal signals. No code change is needed — Atlas Brain resumes routing automatically once the provider recovers. 🔧 For routing misconfigurations: When a model is retired or renamed, engineers update the routing configuration to replace the broken endpoint with an active equivalent. This fix is packaged and deployed as part of a regular engineering release. Until the fix is live, all requests routed to the affected model will continue to fail. 📢 Active incidents affecting Atlas Brain are published on the Search Atlas status page at status.searchatlas.com. Checking this page is the fastest way to confirm whether a known issue is already under active investigation before contacting support. 💡 Key Takeaways for Atlas Agent Users ✅ Intermittent errors are almost always provider-side and self-resolve — wait two to three minutes and retry before taking any other action. ✅ Persistent errors that fire on every request require an engineering fix — clearing your cache or switching browsers will not help. Contact the Search Atlas support team directly so they can identify the affected model and escalate to engineering. ✅ Your account, your data, and your Search Atlas settings are never affected by Atlas Brain routing failures. These errors are fully isolated to the request execution layer and carry no risk of data loss or account impact.

Last updated on Oct 26, 2026

🤖 Atlas Agent: "I'm Temporarily Unable to Process Your Request" — Troubleshooting Guide

When Atlas Agent displays "I'm temporarily unable to process your request," a backend service interruption stopped Atlas Brain from completing your task — nothing you did caused it. Follow the steps below to get back on track within minutes. 🔍 What This Error Means Atlas Brain — the AI engine that powers Atlas Agent — routes your requests to external AI model providers. When a provider is temporarily unavailable due to maintenance, rate limits, or a routing failure, Atlas Brain cannot complete your task and shows this message: "I'm temporarily unable to process your request — this is a hiccup on my end, not anything you did. Please try again in a few minutes. If it's still not working, our support team can help." This error originates entirely on Atlas Brain's side. Your inputs, your account, and your data are not affected. Before troubleshooting, identify which variant you are seeing: - Intermittent — the error recovers within minutes. The self-service steps below will resolve it. - Persistent — the error fires on every request and does not recover after 5–10 minutes. This usually indicates a backend routing misconfiguration. Skip ahead to the support contact section — clearing your cache, switching browsers, or waiting will not help. 🔧 How to Fix the Error Follow these steps in order if the error appears intermittently. Most users are unblocked after step 2. If the error fires on every single request, these steps will not help — go directly to the support contact section below. 1. Wait 2–3 minutes. Most provider outages are brief. Atlas Brain resumes automatically once the provider recovers. 2. Refresh the page and re-submit your request to Atlas Agent. You should see a normal response. 3. Clear your browser cache if the error continues: open your browser settings, clear cached files and cookies for the Search Atlas domain, log back in, then retry. 4. Open an incognito window or switch to a different browser to rule out a local extension or session conflict. After completing these steps, Atlas Agent should respond normally. If the error persists beyond 10 minutes, or if it has been occurring on every request from the start, see the section below. ⏱️ If the Atlas Agent Error Persists Beyond 10 Minutes A prolonged error — especially one that fires on every single request without ever recovering — may indicate a wider service disruption or a backend routing issue that requires an engineering fix. Take these additional steps: - Check the Search Atlas status page at status.searchatlas.com for any active incidents affecting Atlas Agent or Atlas Brain. - Verify your internet connection — an unstable connection can cause requests to fail before they reach Atlas Brain. - Log out and log back in to Search Atlas to refresh your session token, then retry Atlas Agent. If the error remains after completing these steps — or if it was firing on every request from the beginning — contact the Search Atlas support team using the section below. Routing misconfigurations cannot be resolved from the user side and require the engineering team to swap or restore the affected model. 📞 How to Contact Search Atlas Support About the Atlas Agent Error If Atlas Agent still shows the error after you have followed all the steps above — or if the error has been firing on every request from the start, indicating a backend routing issue — the Search Atlas support team can investigate your specific account and session and escalate to engineering if needed. Include the following details when you reach out to speed up the investigation: - Your Search Atlas account email address - The date and approximate time (include your time zone) when the error first appeared - Whether the error occurs on every message you send or only intermittently — this helps the team distinguish a transient provider hiccup from a backend routing issue - A screenshot or copy of the exact error message - The task or prompt you submitted to Atlas Agent when the error occurred - Your browser name and version (for example, Chrome 124 on Windows 11) Reach support via the chat widget inside Search Atlas typing "Human Teammate". The team will confirm receipt and begin investigating right away. 🛠️ Technical Background: What Causes This Error in Atlas Brain Atlas Brain routes your requests to external AI model providers. When a specific model is temporarily unavailable — due to provider-side maintenance, rate limits, or a routing failure on the provider's infrastructure — Atlas Brain cannot fulfil the request and returns the temporary error message to Atlas Agent. In some cases, a specific model may be removed or renamed by the provider, which causes the error to fire on every request until the engineering team swaps in a replacement model on the routing layer. The Search Atlas engineering team actively monitors these failures and deploys fixes as part of regular releases. If you encountered this error, it is very likely already under investigation. 🎯 You now know what triggers the Atlas Agent temporary error and the exact steps to resolve it. If the issue continues after following all the steps above, reach out to Search Atlas support chat, typing "Human Teammate", with your session details and the team will get you unblocked. For a full overview of what Atlas Agent can do, see the Atlas Agent Overview.

Last updated on Oct 26, 2026

🤖 Search Atlas — AI Agent & Atlas Brain Issues

This article covers the AI Agent chat, task execution, tool calls, and Atlas Brain behavior. ⚠️ Error 1: AI Agent shows no response after I send a message What's happening: The agent connection may have dropped or your session needs a refresh. Steps to try: - Check your internet connection — the AI Agent requires a stable, persistent connection. - Refresh the page and try a simple test message ("Hello") to check if the agent is responsive. - Log out and back in to reset your session. - Check remaining AI Agent usage: Top-right Corner Avatar → Billing → See Usage. - Contact support if the agent is completely unresponsive across different browsers. ⚠️ Error 2: Agent starts a task but stops partway through What's happening: The agent may be waiting for clarification, or a specific tool it needed was temporarily unavailable. Steps to try: - Scroll up in the chat — the agent may have asked a question or left a message explaining why it stopped. - Reply with "Continue" or answer any question the agent asked. - Rephrase the request more specifically — "Audit title tags on my homepage" works better than "Fix all SEO issues." - If the agent stopped silently, restate the task in a new message. - Contact support with the conversation transcript if the agent consistently stops at the same point. ⚠️ Error 3: Agent recommendations reference issues I already fixed What's happening: The agent is using data from a previous OTTO crawl that predates your fixes. Steps to try: - Trigger an OTTO re-crawl: Left Sidebar → OTTO SEO → Site Audit → Overview (Website Overview) → Recrawl Site. - After the crawl completes, start a new AI Agent conversation — the agent loads fresh data at session start. - Tell the agent directly: "I've fixed [issue]. Please check the current crawl data." - Mark resolved tasks as complete in the OTTO task list so they don't resurface. - Confirm your fixes are live on the published URL, not just saved in drafts. ⚠️ Error 4: Agent says it performed an action but nothing changed What's happening: Some actions are queued rather than executed immediately, or a required integration is missing. Steps to try: - Check the OTTO task list for any pending approvals generated by the agent. - Hard-refresh the relevant module (Content Assistant, OTTO, etc.) to see if changes are visible there. - Ask the agent: "Please confirm the exact action you took and where I can verify it." - Check that the required integration is active — some actions require a connected CMS or GSC account. - Contact support with the specific action the agent claimed to perform and where you expected to see it. ⚠️ Error 5: Agent responses are too generic and not specific to my site What's happening: The agent needs more context — either from your prompt or from Brand Vault. Steps to try: - Be specific in every message — include your domain, target keyword, and specific goal. - Complete your Brand Vault fully: More Features → Brand Vault. - Ask data-driven questions: "Based on my current OTTO audit, what are my top 3 SEO priorities?" - Include your goals: "My goal is 20% more organic traffic in 6 months. Where should I focus?" - Test what context the agent has: "Describe my business as you understand it" — the answer reveals how much Brand Vault context is being used. ⚠️ Error 6: Agent isn't using my Brand Vault information What's happening: Brand Vault may not be linked to the current project, or Brand Vault is incomplete. Steps to try: - Verify Brand Vault has content in every section: More Features → Brand Vault. - Link Brand Vault to your project: OTTO SEO in the left sidebar → All Sites → open your project → Brand Vault. - Start a new chat session — Brand Vault context loads at session start. - Explicitly reference it, even you can type the Vault name. - Contact support if the agent returns no Brand Vault context after completing all sections. ⚠️ Error 7: Agent can't access a tool it's trying to use What's happening: The specific tool's underlying service (Keyword Research, Content Assistant, GSC) was temporarily unavailable. Steps to try: - Wait 5 minutes and ask the agent to try again. - Check if the relevant module works on its own in the dashboard. - Ask the agent: "Which tool failed? Can you try a different approach?" - Ask the agent to skip the failing step and continue with the rest of the task. - Contact support with the specific task and which step failed. ⚠️ Error 8: Complex tasks time out before completing What's happening: Multi-step agent workflows that take more than a few minutes can exceed connection timeouts. Steps to try: - Break large tasks into smaller sequential requests across separate messages. - Check if the agent left partial results — often the first few steps completed before the timeout. - Use the agent's checkpointing: "Continue from where you left off" in a new message. - For long-running jobs, ask the agent to provide a status update every few steps. - Contact support if critical multi-step tasks consistently time out at the same step. The AI Agent is not a static tool—it’s a dynamic system that depends on context, data freshness, integrations, and task clarity. Most issues arise not from failure, but from missing context, outdated data, or overly broad instructions.

Last updated on Oct 26, 2026

🛠️ Troubleshooting Search Atlas Coworker in Slack

If Atlas isn't behaving as expected, this guide walks through the most common issues and how to fix them. Work through the steps for your situation — and if you're still stuck, you can reach a person any time (see the last section). 💬 Atlas Isn't Responding or Posting in a Channel 1. Make sure it's a regular channel. Atlas can't post in Slack Connect or externally-shared channels — try a regular public or private channel, or a direct message instead. 2. Invite Atlas to the channel by typing /invite @Search Atlas Coworker and selecting the bot from Slack's autocomplete. 3. Confirm you're a member of the channel — Atlas only works in channels you belong to. 4. To isolate the issue, try the same request in a direct message with Atlas. If it works there, the problem is specific to the channel. 🔗 Problems Connecting or Using an App You connected an app but Atlas says it isn't connected. Ask Atlas again — the connection sometimes needs a moment to register. If it persists, ask Atlas to reconnect the app in plain language (for example, "reconnect my HubSpot") or manage the connection from the Atlas app's Home tab in Slack. 1. The connect link shows a "pipedream.com" address. This is expected and safe — it's the secure sign-in step. Go ahead and approve the connection. 2. You're connecting several apps. Connect them one at a time rather than asking for several in a single message. 3. All your connected apps suddenly stop working (but Search Atlas tools still work). This is usually a temporary connection-service hiccup. Wait 5–10 minutes and try again. ✅ Atlas Isn't Completing an Action (Calendar Event, Publishing, etc.) 1. If your workspace is configured to require approval, Atlas shows an approval card before completing the action. If you expected a confirmation prompt but didn't see one, check Workspace Settings → Permissions → Run connector actions without asking in the Search Atlas Coworker dashboard to verify approval is turned on for that connector or tool. 2. After it runs, verify the result in the app itself (for example, check your calendar or your CMS) to confirm it went through. 3. If it still doesn't complete, try the request again. 4. Publishing to WordPress: both WordPress.com and self-hosted WordPress are supported. Make sure the connection is authorized (for self-hosted, this may use an application password). 🔄 A Routine Ran Multiple Times or Didn't Run 1. Ask Atlas to "pause all routines." 2. Delete the affected routine. 3. Recreate the routine you want, describing the task and schedule again. 💡 Atlas's Answers or Behavior Seem Off 1. An action worked once and failed the next time. Try it again — many of these are temporary. If it keeps failing for the same action, reach out to us. 2. Atlas says Slack isn't connected when you ask it to summarize a channel. Rephrase the request explicitly, for example: "Use my Slack connection to summarize #channel-name." 3. A very long answer looks garbled. Ask Atlas for a shorter version, e.g. "Can you give me a shorter version?" 4. Atlas refers to the wrong person. Make sure you've signed in with /atlas-login. 5. A button's label looks cut off. You can type your choice directly in the message box instead of clicking the button — Atlas accepts typed responses. ⏳ Atlas Seems Slow or Unresponsive While Atlas is working, it shows an hourglass on its own message. Complex requests — like a full site audit — can take up to around 90 seconds, so give it a moment. If nothing comes back after a few minutes, try the request again. ⌨️ A Slash Command Shows as Plain Text and Nothing Happens Type the command directly in the main Slack message box, not as a reply inside a thread. If it still shows as plain text, retype it in the Slack app. 💬 Still Need Help? You can reach us through live chat right here. If you'd like to talk to a person, just type human teammate in the chat and we'll connect you with one of our team.

Last updated on Oct 26, 2026

🔌 Troubleshooting 'Failed to Fetch' AI Agent Errors

This article addresses the 'failed to fetch' error that appears when working with the AI agent inside Content Genius — specifically when the agent fails to respond or load results regardless of what you type. This error is a connection-level failure, not a problem with your input, and it has specific causes and fixes you can work through step by step. 🔍 Common Causes - Browser or network interruption: The AI agent loses its connection mid-session, often due to a timeout, unstable Wi-Fi, or a VPN interfering with requests. - Stale session or cached data: An expired session token or corrupted cache can block the agent from making successful API calls. - Browser extension conflict: Ad blockers, privacy extensions, or script blockers frequently intercept AI agent fetch requests. - Temporary platform instability: A brief backend outage or degraded service can produce this error across all users intermittently. 📍 Where to Find It The AI agent is accessed via Left sidebar → Home (Hi, CSM. What will we get done today?). Once there, click Launch Atlas Agent as a per-project action to open the agent interface where the error appears. 🛠️ Step-by-Step 1. Reload the page: Press Ctrl + Shift + R (Windows) or Cmd + Shift + R (Mac) to do a hard refresh, bypassing the cache. Then relaunch the agent via Left sidebar → Home (Hi, CSM. What will we get done today?) → Launch Atlas Agent (per-project action). 2. Check your internet connection: Open another tab and load a different website to confirm your connection is stable. If you are on Wi-Fi, try switching to a wired connection or a mobile hotspot to rule out network instability. 3. Disable browser extensions: Open an Incognito or Private window (which disables most extensions by default) and navigate back to Left sidebar → Home (Hi, CSM. What will we get done today?) to test whether the error persists. If the agent works in Incognito, a browser extension is the cause — disable them one by one to identify the culprit. 4. Clear browser cache and cookies: In your browser settings, clear cached images, files, and cookies for the Search Atlas domain, then log back in and retry the agent. 5. Try a different browser: If the error continues, open Search Atlas in a different browser (e.g., switch from Chrome to Firefox or Edge) and relaunch the agent to isolate whether the issue is browser-specific. 6. Disable your VPN (if applicable): If you are using a VPN, temporarily disconnect it and retry. VPNs can route traffic in ways that block AI agent fetch requests. 7. Wait and retry: If none of the above steps resolve the issue, wait 5–10 minutes and try again, as this error can also be caused by a temporary platform-side issue that resolves on its own. ✅ How to Confirm It Worked After applying any of the steps above, confirm the fix by doing the following: - Navigate to Left sidebar → Home (Hi, CSM. What will we get done today?) and click Launch Atlas Agent (per-project action). - Type any prompt into the agent input field and submit it. - A successful connection will show the agent actively processing your request — you will see a loading indicator followed by a response, not the 'failed to fetch' error message. - If the agent responds — even partially — the connection has been restored. If the error reappears immediately, note the exact steps you already tried and contact support using the method below. If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

⚠️ Cloud Stack Shows a Blank Preview After Generation Completes

If your Cloud Stack preview is blank even though the orchestrator has confirmed generation is complete, a known bug may have falsely reported success before your content finished building. Trigger a fresh regeneration using the steps below — your settings, keywords, and configuration are preserved. ⚠️ What Causes the Blank Preview The Cloud Stack orchestrator has a known issue where it reports "fully generated and ready for review" without first confirming that content is actually available. This false-success state most often occurs after multiple edit or regenerate cycles on the same stack. The preview page is blank because the underlying content did not finish building — the confirmation message is inaccurate. 🔍 How to Confirm This Is the Issue Look for all three of these signs before following the recovery steps: - The orchestrator or agent shows a "fully generated" or "ready for review" success message. - The Cloud Stack preview page loads completely blank — no sections, headings, or body content appear. - The blank preview occurred after at least one prior edit or regenerate cycle on the same stack. 🛠️ How to Recover a Blank Cloud Stack Preview Triggering a single fresh regeneration resolves the issue in most cases. 1. Go to Authority → Cloud Stacks. 2. Locate the affected stack in your project list. 3. Click Regenerate. Do not create a new stack — regenerating keeps all saved settings. 4. Stay on the page until generation fully completes. Do not navigate away or refresh mid-generation. 5. Open the Preview tab and confirm that sections and body content are visible before publishing or scheduling. The fix is successful when the Preview tab displays your stack's full content. ⚠️ What to Avoid During Recovery - Do not click Regenerate multiple times in quick succession — repeated triggers can reproduce the same false-success state. - Do not publish or schedule the stack until the Preview tab confirms visible content. - If the preview is still blank after regeneration completes, wait 2–3 minutes, refresh the page, then try once more. 🔑 When to Contact Support If the preview remains blank after two separate regeneration attempts, contact the Search Atlas support team. Include the following in your message: - The stack name and the project it belongs to. - A screenshot of the blank preview page. - A screenshot of the "fully generated" confirmation message from the orchestrator. The support team can inspect the stack's generation state and reset it directly. 🎯 You now know how to identify and recover from the Cloud Stack blank-preview bug — regenerate the affected stack and confirm visible content in the Preview tab before publishing. If two regeneration attempts fail, contact Search Atlas support with the details above.

Last updated on Oct 26, 2026

⚠️ Atlas AI Agent: NLP Terms Not Fully Deploying

The Atlas AI Agent may stop after the first batch of NLP terms — or report "no deployable fixes" — before finishing the full deployment. Re-prompting the agent with your original request will complete the process. 🔍 What's Happening The Atlas AI Agent processes NLP terms in batches. It may stop before completing all batches in a single session. Symptoms you may notice: - The agent reports "no deployable fixes" even when NLP terms are clearly pending. - The agent deploys only the first batch of NLP terms, then stops. - You need to re-prompt the agent multiple times before all NLP terms are deployed. This affects you if you're using the Atlas AI Agent to deploy NLP terms in bulk or in batches. 🛠️ Workaround Steps Follow these steps in the Atlas AI Agent chat to complete the full NLP terms deployment: 1. Open the Atlas AI Agent chat and send your original NLP terms deployment request. 2. Wait for the first batch to finish deploying. 3. Re-send the same deployment request after each batch completes. 4. If the agent replies with "no deployable fixes," dismiss the message and prompt it again with the same original request. 5. Repeat steps 3 and 4 until the agent confirms that all NLP terms have been deployed. You'll know the deployment is complete when the Atlas AI Agent confirms there are no remaining NLP terms to deploy. 📡 Engineering Status Our team is aware of this issue and is actively investigating a permanent fix so the Atlas AI Agent can complete NLP terms deployments in a single session without requiring re-prompts. Until the fix ships, the workaround above will get your full NLP terms list deployed successfully. We apologize for the inconvenience. 📞 When to Contact Support Reach out to Search Atlas support if: - You've repeated the workaround multiple times and NLP terms are still not deploying. - The Atlas AI Agent stops responding or shows an error other than "no deployable fixes." - You're unsure whether your NLP terms deployed successfully. Send a message through the in-app chat and include a brief description of what you've tried so the support team can assist you faster. 🌀 The re-prompt workaround will get your full NLP terms list deployed while our engineering team works on a permanent fix — if you continue to hit issues, contact Support through the in-app chat.

Last updated on Oct 26, 2026

🛠️ Fix a Persistent Content Assistant Loading Error

🔍 Overview Some users encounter a persistent "Error loading Content Assistant" message when trying to open the editor. Unlike a typical temporary glitch, this error survives hard refreshes, browser cache clears, incognito mode, and even affects other users on the same account — such as a virtual assistant. This article explains what causes it, what you can do on your end, and when to contact support. ⚠️ Recognising This Error This issue is different from a standard loading hiccup. Look for these signs: - The Content Assistant editor fails to load every time, not just occasionally. - The error persists across multiple browsers and devices. - Logging out and back in does not fix it. - Clearing your browser cache and cookies has no effect. - Opening the editor in an incognito or private window still shows the error. - Other team members or virtual assistants on your account are blocked too. - The problem has lasted more than a few hours or several days. 🧭 Steps to Try First Before reaching out to support, work through these steps in order: 1. Hard refresh the page. Press Ctrl + Shift + R (Windows) or Cmd + Shift + R (Mac) to bypass any locally cached version of the page. 2. Clear your browser cache and cookies and then reload the Content Assistant editor. 3. Try a different browser. If you are using Chrome, test in Firefox or Edge, and vice versa. 4. Open an incognito or private browsing window and navigate to the editor again. 5. Log out of Search Atlas completely, close all tabs, then log back in and retry. 6. Try from a different device or network to rule out a local connection issue. If none of the above steps resolves the error, the root cause is almost certainly a backend issue affecting your account — not something on your device or network. 💡 What Actually Fixes This Based on confirmed support cases, this type of persistent Content Assistant error is caused by a backend problem that cannot be resolved through any client-side action. When this happens: - No amount of refreshing, cache clearing, or browser switching will fix it on your end. - The fix requires a backend correction applied by the Search Atlas engineering or support team. - Once the fix is applied, the Content Assistant editor loads normally for all users on the account, including virtual assistants. This is why it is important to contact support promptly if you have gone through all the steps above and the error continues. The sooner the team is notified, the sooner a backend fix can be applied. 📋 Information to Have Ready When Contacting Support To speed up resolution, gather the following before opening a chat: - The URL of the Content Assistant editor page that is showing the error. - Approximately when the error first appeared. - Confirmation of which troubleshooting steps you have already tried (browsers tested, cache cleared, incognito, etc.). - Whether other users or team members on your account are also affected. - Any screenshot or screen recording of the exact error message if possible. 🚀 After the Fix Is Applied Once the backend issue has been resolved by the support team, the Content Assistant editor should load correctly immediately. You do not need to take any additional action on your end. Simply navigate back to the editor and confirm it is working. If the error returns after being resolved, contact support again right away with the details. 💬 Get Help Now If you have completed all the troubleshooting steps above and the Content Assistant editor is still not loading, this requires a backend fix from our team. If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

🛠️ AI Agent Orchestrator: Website Building Routes to Wrong Tool

When you ask the AI Agent to build a website, it should open Website Studio automatically. In some cases, the agent opens Cloud Stack instead — or skips the action entirely when your request is phrased in past tense. ⚙️ Why the AI Agent Selects the Wrong Tool The AI Agent Orchestrator reads the phrasing of your request to decide which tool to open. Two patterns trigger incorrect routing: - Cloud Stack opens instead of Website Studio: Certain keyword combinations in your prompt match the Cloud Stack routing rule before the Website Studio rule is evaluated. - No action is taken: Past-tense phrasing — for example, "I built a website" or "I was creating a site" — signals to the agent that the task is already complete, so it skips launching any tool. 💡 How to Phrase Website Building Requests Correctly Use present-tense, action-oriented language so the AI Agent routes to Website Studio. - Use: "Build a new website" or "Create a website" - Use: "Open Website Studio" or "Start a new site in Website Studio" - Avoid: "I built a website" or "I was building a site" — past tense causes the agent to treat the task as done - Avoid: vague phrases like "work on my site" without specifying the action you want taken Naming the tool explicitly — for example, "Use Website Studio to create a site" — ensures correct routing regardless of other wording in your message. 🛠️ What to Do If Cloud Stack Opens Instead of Website Studio 1. Close or dismiss the Cloud Stack panel that opened. 2. Return to the AI Agent chat. 3. Re-enter your request using present-tense phrasing — for example: "Build a new website in Website Studio." Website Studio will open on your next prompt and you can begin building immediately. ⚠️ Known Issue Status This routing behavior has been identified and resolved by the Search Atlas engineering team. If you are still seeing Cloud Stack open for website building requests after rephrasing your prompt, contact Search Atlas support so the team can investigate your specific account. 🎯 You now know why the AI Agent may open Cloud Stack instead of Website Studio and how present-tense phrasing gets you to the right tool every time. If the issue persists after rephrasing, reach out to Search Atlas support for hands-on assistance.

Last updated on Oct 26, 2026

🛠️ Fix AI Agent Technical Errors Quickly

🔍 What Is This Error? If you see a message like "Sorry, there was an error processing your message" or the AI Agent stops responding mid-task, you have encountered a temporary technical error. These errors can occur when the agent tries to process a request, apply on-page fixes, or retrieve content details. The good news: most of these errors are transient and resolve quickly with a few simple steps. ⚠️ Common Situations Where This Happens - The AI Agent returns an error immediately after you send a message. - The agent starts a task — such as deploying on-page fixes — and then stops with an error. - Post details or content previews fail to load and show a technical error instead. - The agent appears to abandon your request without completing it. 🚀 How to Fix the Error 1. Refresh the page. Open OTTO SEO → All Sites (SEO Automation) from the left sidebar and reload your browser tab. This clears temporary session issues that can interrupt agent tasks. 2. Retype your request. Start a new message to the AI Agent. Do not copy and paste the previous message — rephrase it slightly and try again. 3. Wait one to two minutes. Occasional spikes in platform load can cause brief errors. Waiting a short time before retrying usually resolves this. 4. Check your browser. Clear your browser cache or try a different browser (Chrome or Edge recommended). Browser extensions can sometimes interfere with the agent. 5. Try a different task scope. If the error happens on a specific page or fix, try applying the action to a different URL first to confirm the issue is not isolated to one piece of content. 🔄 Why These Errors Occur AI Agent errors are typically caused by one of the following: - Temporary server timeouts — the agent's request takes longer than expected and the connection resets. - Tool availability changes — occasionally, an internal tool the agent relies on is updated or temporarily unavailable, causing the agent to stop mid-task. - Session expiry — long inactive sessions can cause the agent to lose context, resulting in a failed response. Our engineering team actively monitors error patterns and deploys fixes rapidly. Known issues related to agent errors have already been resolved as part of ongoing platform improvements. ✅ What to Do If the Error Persists If you have followed the steps above and the AI Agent continues to return errors, gather the following information before reaching out to support — it helps our team investigate faster: - The exact error message you see on screen. - The task or request you were trying to complete. - The URL or page the agent was working on. - The approximate time the error occurred. If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

🔍 Fix an Empty Domain Knowledge Network

🧩 What Is the Domain Knowledge Network? The Domain Knowledge Network is a visual map that shows how your business entity connects to related topics, people, places, and concepts across the web. It is powered by the data you enter in your Knowledge Graph, but it relies on additional processing steps before it populates. ⚠️ Why It May Appear Empty A common point of confusion is assuming that completing 100% of the Knowledge Graph fields will immediately fill the Domain Knowledge Network. In reality, there are a few reasons why the network may still appear empty even after full completion: - Processing delay: After your Knowledge Graph reaches 100% completion, the platform needs time to process and build the entity connections. This can take up to 24–48 hours. - Entity recognition pending: The system must verify and match your business entity against external knowledge bases. If your business is newer or has limited online presence, this step may take longer. - Missing or inconsistent NAP data: Your business Name, Address, and Phone number (NAP) must be consistent and complete. Even small discrepancies can prevent entity matching from completing successfully. - Schema or website data not yet crawled: If your website was recently updated with structured data or business schema, search engines and the platform may not have re-crawled it yet. ✅ Steps to Resolve the Issue 1. Confirm 100% Knowledge Graph completion. Navigate to your Knowledge Graph dashboard and verify every section — including business categories, social profiles, founding details, and key people — is fully filled in. Partial entries in optional fields can still affect processing. 2. Check your NAP consistency. Make sure your business name, address, and phone number exactly match what appears on your website, Google Business Profile, and other online listings. Even minor differences (e.g., "St." vs. "Street") can cause issues. 3. Verify your website URL is correctly entered. An incorrect or redirecting URL can prevent the platform from crawling your site and confirming your entity data. 4. Wait 24–48 hours and refresh. If everything looks correct, allow up to two full days for the Domain Knowledge Network to populate. Refresh the page after this period to check for updates. 5. Save and re-submit your Knowledge Graph. Try opening your Knowledge Graph, making a minor edit to any field, saving the changes, then reverting and saving again. This can trigger a re-processing of your entity data. 💡 Tips to Improve Entity Recognition The strength and accuracy of your Domain Knowledge Network depends on how well your business entity is established online. Consider the following best practices: - Ensure your business has a Google Business Profile that is fully verified and up to date. - Add structured data (Schema.org markup) to your website, especially LocalBusiness or Organization schema types. - Maintain consistent business listings across major directories such as Yelp, Bing Places, and industry-specific platforms. - Link your social media profiles (LinkedIn, Facebook, Twitter/X) within the Knowledge Graph to strengthen entity signals. 🔄 When to Expect Results Most customers see their Domain Knowledge Network begin to populate within 24–48 hours of completing a fully consistent Knowledge Graph. Businesses with a stronger existing online presence tend to see faster results. If the network remains empty after 48 hours and you have followed all the steps above, it is worth reaching out to our team for a manual review. 🛠️ Still Seeing an Empty Network? If your Domain Knowledge Network is still empty after 48 hours and you have verified all the steps in this article, our team can investigate your specific entity data and processing status. If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

🔧 Fix Stalling Approvals in Slack and Dashboard

Overview Some users have experienced a frustrating issue where approval actions sent through Slack or the Search Atlas Dashboard appear to stall indefinitely — hitting approve has no effect, and the workflow never moves forward. This article explains what caused the issue, how it was resolved, and what to do if you encounter similar behavior. What Caused This Issue This problem was traced to a confirmed engineering bug affecting the approval flow for AI agent and Coworker workflows. The bug caused approval actions to fail silently: the system appeared to accept your approval, but nothing executed behind the scenes. No status update was sent to Slack or the Dashboard, leaving the workflow stuck with no indication of what went wrong. This issue has since been identified and resolved by the engineering team. Current Status The engineering team has completed the fix. If you were affected by this bug, approval actions through both Slack and the Search Atlas Dashboard should now work reliably. No action is required on your end to apply the fix — it was deployed automatically. Steps to Verify Everything Is Working 1. Open the Search Atlas platform and navigate to your AI agent or Coworker workflow area. 2. Trigger or locate a pending approval action in your workflow. 3. Submit your approval and wait a moment for the system to respond. 4. Check whether the workflow advances as expected — either in the Dashboard or in your connected Slack channel. 5. If the action completes successfully, no further steps are needed. If You Are Still Experiencing Stalling If approval actions are still not moving forward after the fix, try the following before contacting support: - Refresh your browser and retry the approval. Occasionally a stale session can cause the interface to appear unresponsive even when the backend is functioning. - Clear your browser cache and reload the platform, then attempt the approval again. - Check your Slack connection within the Search Atlas Dashboard to confirm your Slack integration is still active. If it appears disconnected, use the available option to reconnect it. - Try approving from a different surface — if the action is stalling in Slack, attempt it directly from the Search Atlas Dashboard, or vice versa. - Wait a few minutes and retry. Transient network or service interruptions can occasionally delay responses. If the action processes after a short wait, no further action is needed. What to Include If You Need Further Help If you are still experiencing stalling after trying the steps above, please reach out to our support team. Include your account details, a description of the workflow involved, and the surface where the stalling is occurring (Slack, Dashboard, or both). If you need assistance, you can contact us through the Search Atlas support channel.

Last updated on Oct 26, 2026

🛠️ Fix Atlas Agent Photo Count Restriction Errors

🔍 What Is Happening When you launch Atlas Agent to automatically create content such as Press Releases or Cloud Stacks, the Agent pulls all images stored in your Brand Vault and attempts to attach them to the content it is building. If the total number of images in your Brand Vault exceeds the limit the automated system can attach to a single piece of content, the Agent stops and returns a photo count restriction error. This is a known behaviour in the automated pipeline. Your Brand Vault data is safe and nothing has been deleted or corrupted. The error only affects automated content generation — it does not reflect a problem with your account or your images themselves. ✅ Why Manual Creation Still Works When you create a Press Release or Cloud Stack manually, you choose which images to attach at the time of creation. Because you are selecting a smaller set of images rather than attaching the entire Brand Vault automatically, the photo count limit is never reached and the content is created successfully. ⚙️ How to Work Around This Issue Until a permanent platform fix is fully rolled out, the most reliable workaround is to reduce the number of images in your Brand Vault so it stays within the limit the Atlas Agent can handle automatically. Follow the steps below: 1. Audit your Brand Vault images. Review all images currently saved in your Brand Vault and identify any that are duplicates, outdated, or not needed for automated content generation. 2. Remove excess images. Delete or archive images you do not need the Agent to use automatically, keeping only the core brand images you want attached to generated content. 3. Retry Atlas Agent. Once the image count is reduced, go to Left sidebar → Content → Content Genius and relaunch Atlas Agent to generate your Press Release or Cloud Stack. 4. Use manual creation as an immediate alternative. If you need content right away and cannot reduce your Brand Vault images immediately, use the manual creation flow inside Left sidebar → Content → Content Genius to select only the images you need and complete your content without hitting the restriction. 📋 Key Facts to Know - Your Brand Vault images and all account data remain completely safe — this error does not cause any data loss. - The photo count restriction is triggered by the total number of images in your Brand Vault, not by the size or format of the images. - Manual content creation is a fully supported alternative and produces the same output as the automated Agent flow. - Engineering has addressed related Brand Vault sync issues in recent platform updates, and improvements to the automated image-handling pipeline are ongoing. 💡 Tips to Prevent This in the Future - Keep your Brand Vault lean by storing only the images you actively use for content generation. - Regularly review and remove any images that were uploaded for one-off projects and are no longer needed. - If you add a large batch of images to your Brand Vault, test Atlas Agent on a small content task immediately afterwards to confirm the count is still within limits before running bulk generation. 🙋 Still Seeing the Error If you have reduced your Brand Vault images and Atlas Agent is still returning a photo count restriction error, or if you are unsure how many images are within the allowed limit, our team can investigate your specific account. If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

🔧 Fixing Atlas Agent 400 Errors and Keyword Data

Overview If you are seeing a Request failed: 400 error while using the Atlas Agent on content tasks, or if your keyword data is not loading correctly, this article will walk you through the most common causes and how to resolve them quickly. These errors are typically caused by expired sessions, stale cache data, API timeouts, or a temporary backend issue — none of which reflect a problem with your account or subscription. What Does a 400 Error Mean? A 400 (Bad Request) error occurs when the platform cannot process the request sent by your browser or session. In the context of the Atlas Agent and content analysis tools, this usually means one of the following: - Your session has expired or become invalid mid-task - The browser has cached outdated authentication or request data - A temporary backend service interruption is affecting API responses - A request timed out before the server could respond correctly - Your SSO (Single Sign-On) token has expired and needs to be refreshed Step-by-Step Troubleshooting Follow these steps in order. Most customers resolve the issue within the first two or three steps. 1. Refresh the page. Reload the page in your browser and retry your Atlas Agent task or keyword data request. This resolves temporary communication errors caused by brief API timeouts. 2. Clear your browser cache and cookies. Stale cached data is a leading cause of 400 errors. Access your browser's settings and clear cached files and cookies, then reload the platform and sign back in. The exact steps vary by browser — refer to your browser's help documentation if needed. 3. Sign out and sign back in. Go to your Avatar Menu (👤) and select Sign Out. Wait 30 seconds, then sign back in. This refreshes your session token and resolves the majority of authentication-related 400 errors. 4. Re-authenticate via SSO if applicable. If your organisation uses Single Sign-On, your SSO token may have expired without triggering a visible logout. Sign out completely, then log back in through your SSO provider to obtain a fresh token before returning to Search Atlas. 5. Try a different browser or incognito window. Open an incognito or private window (or a different browser entirely) and sign in to Search Atlas. If the Atlas Agent and keyword data load correctly in this context, the issue is isolated to a browser extension, cached data, or a profile-specific setting in your main browser. 6. Disable browser extensions. Ad blockers, privacy tools, and script blockers can interfere with API requests. Temporarily disable extensions, reload the platform, and retry your task. If Keyword Data or SEO Rankings Are Still Not Showing If the 400 error clears but your keyword data or SEO ranking informa If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

🔍 Fix Atlas Brain Hallucinations and OTTO Ads Issues

🔍 What Is an Atlas Brain Hallucination? Atlas Brain is Search Atlas's AI assistant. Like all large language models, it can occasionally generate content that looks accurate but contains fabricated details — such as pages, reports, or campaign data that do not actually exist on your site. This is known as a hallucination. Common examples include: - A website redesign report referencing pages that were never created - Keyword or traffic data that does not match your actual analytics - Campaign summaries describing ad groups or assets that do not exist in your account Atlas Brain hallucinations are a known AI limitation. They do not reflect changes made to your site or ad accounts — no content is published, no pages are created, and no settings are changed without your explicit action. ⚙️ Why Your OTTO Ads Campaigns May Be Paused OTTO Ads campaigns can pause when the Google Ads account linked to OTTO becomes disconnected or mismatched. When the Google Ads account is no longer properly connected to OTTO, OTTO Ads will automatically pause affected campaigns to prevent uncontrolled spend. If your campaigns are paused, check whether your Google Ads account is still correctly connected to the platform. A mismatch or disconnection between the account linked in Search Atlas and your actual Google Ads account is the most common root cause. 💳 Are You Still Being Charged While Campaigns Are Paused? Your Search Atlas subscription fee covers access to the platform and its tools — it is not Google Ads spend. These are two separate charges: - Search Atlas subscription: Billed by Search Atlas on your regular billing cycle, regardless of whether campaigns are active or paused. - Google Ads spend: Billed directly by Google. When OTTO Ads campaigns are paused, Google stops serving those ads and you stop accruing ad spend with Google for those campaigns. You will not be charged Google Ads spend for paused campaigns. If you see unexpected charges on your Google Ads account, those originate from Google and should be reviewed in your Google Ads billing dashboard. 🛠️ How to Reconnect Your Google Ads Account To restore your OTTO Ads campaigns, ensure the correct Google Ads account is connected to the platform. Navigate to your account's integrations or connected accounts area and verify that the Google Ads account shown matches the account you intend to use with OTTO Ads. If the connection appears broken or points to the wrong account, disconnect it and reconnect the correct Google Ads account. Once the correct account is linked, your OTTO Ads campaigns should be able to resume. If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

🛠️ Fix OTTO AI Not Following Custom Prompt Instructions

🔍 Overview When using OTTO AI to generate content, you can provide custom prompt instructions to control tone, style, and specific rules — such as avoiding your company name or keeping output within a certain length. If OTTO appears to ignore these instructions, a few common causes are usually responsible. This article explains what to check and how to fix it. ⚠️ Common Symptoms - The AI includes your company name even after you instructed it not to. - Generated content is cut off or truncated unexpectedly. - Output does not reflect the tone, format, or restrictions you specified. - Instructions seem to apply inconsistently across different generations. 💡 Why This Happens OTTO AI processes your custom instructions alongside its built-in SEO logic. In some cases, conflicts between these two sets of rules can cause certain instructions to be deprioritised or missed entirely. The most frequent causes are: - Instruction placement: Instructions buried in long paragraphs are harder for the model to parse reliably. - Conflicting directives: Asking the AI to both maximise content length and avoid repeating certain words can create tension that leads to unpredictable output. - Vague phrasing: Ambiguous instructions such as "keep it short" leave room for interpretation. Specific word or character limits work better. - Token limits: Very long custom prompts combined with large content requests can push the total input near its processing limit, causing instructions near the end to be dropped or content to be cut off. ✅ How to Write Effective Custom Prompts Follow these best practices to ensure OTTO AI respects your instructions every time. 1. Put critical rules first. Place your most important instructions at the very beginning of the custom prompt field, before any stylistic guidance. 2. Be explicit and specific. Instead of "don't mention the company name," write: "Do not use the word [Company Name] anywhere in the output." 3. Set a clear length target. Specify an exact word count range, for example: "Write between 400 and 500 words." Avoid open-ended instructions like "keep it brief." 4. Use a list format for rules. Numbered or bulleted rules inside your prompt are easier for the model to follow than rules written as flowing prose. 5. Avoid contradictory instructions. Review your prompt for conflicting requirements before generating. For example, do not ask for comprehensive detail and a strict 200-word limit at the same time. 6. Keep the prompt concise. Aim for under 200 words in your custom instructions. Longer prompts increase the risk of critical rules being deprioritised. 🚀 Step-by-Step: Updating Your Custom Prompt in OTTO AI 1. Go to Left sidebar → OTTO SEO → SEO Automation (URL: /seo-automation-v3). 2. Open the OTTO AI generation settings for the relevant page or task. 3. Locate the Custom Instructions or Custom Prompt field. 4. Clear the existing instructions and rewrite them using the best practices above. 5. Place your exclusion rules (e.g., "Do not use [Company Name]") on the first line. 6. Add a specific word count target on the second line. 7. Save your settings and run the generation again. 8. Review the output to confirm your instructions are being followed before applying it to additional pages. 📊 Additional Tips for Consistent Results - Test with a single page first. Validate that your revised prompt works correctly before applying it in bulk. - Use the AI Content Templates tool for recurring content types. Navigate to Left sidebar → Content → AI Content Templates to save standardised prompts so you do not need to re-enter rules each time. - Check output length settings. If content is being cut off, verify that no conflicting length cap is set inside the generation options panel, separate from your custom prompt. - Regenerate if needed. If a single generation still does not comply, regenerate once more. Occasional variation is normal; persistent non-compliance points to a prompt issue. 🙋 Still Need Help? If you have followed all the steps above and OTTO AI is still not following your custom instructions, If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

🔧 Troubleshooting Atlas Brain Fetch Errors

If you are seeing a 'failed to fetch' error in Atlas Brain, this article covers general troubleshooting steps you can take and what information to collect if the issue needs to be escalated to our team. General Troubleshooting Steps 1. Refresh the page. A simple page refresh can resolve temporary connection issues between your browser and the Atlas Brain service. 2. Clear your browser cache and cookies. Stale browser data can sometimes interfere with requests. Clear your cache, reload the platform, and try your prompt again. 3. Try a different browser or incognito mode. This helps rule out browser extensions or cached sessions as the cause. 4. Test with a simple prompt. Enter a basic prompt (e.g., 'Summarize my latest project') to see whether the error occurs regardless of prompt complexity. 5. Check your internet connection. Intermittent connectivity can cause fetch errors. Confirm your connection is stable before retrying. 6. Log out and back in. Signing out of the platform and signing back in can refresh your session and resolve authentication-related fetch errors. If the Error Persists If the 'failed to fetch' error continues after trying the steps above, the issue may require investigation by our engineering team. Please have the following information ready when you reach out so we can diagnose the problem quickly: - Exact error message — the full text of the error you see, including any codes or details. - Timestamps — when the error started and whether it happens every time or only intermittently. - Prompt details — the type of prompt or input you were using when the error appeared. - Browser information — your browser name and version, and whether you tried other browsers. - Project context — the project name or workspace where the error is occurring. - Recent changes — any recent changes to your account, projects, or connected data sources. Providing these details upfront will help our team reproduce the issue and identify a root cause faster. If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

🔄 Recovering AI Credits After Atlas Brain Overconsumption

🔍 What Happened Atlas Brain, the AI agent inside Search Atlas, runs tasks autonomously to complete your goals. When you set a credit limit — for example, asking it to generate up to 1,000 Twitter cards — it may interpret the task broadly and perform additional actions beyond what you expected. In some cases, this causes the agent to consume your entire AI credit quota before you can intervene. This article explains why this happens, how to reduce the risk of it recurring, and what to do if you need your quota reset. ⚠️ Why Atlas Brain May Exceed Your Intended Credit Limit Atlas Brain is designed to complete goals, not just individual actions. If you phrase a request in a way that implies a broader outcome, the agent may generate related content or run additional steps it considers necessary — even if you intended a narrower scope. The credit ceiling you mention is treated as a budget cap, but the agent may reach that cap by completing tasks you did not explicitly anticipate. Key things to be aware of: - Specifying a credit amount does not restrict the type of actions Atlas Brain can take — only the total spend. - The agent may combine multiple task types to fulfil what it interprets as your goal. - Once credits are fully consumed, the agent stops, but the quota is already spent. ✅ How to Reduce the Risk of Overconsumption Before running Atlas Brain on a large task, follow these best practices: 1. Be specific about scope. Instead of open-ended goals, describe exactly what you want — for example, "Generate Twitter card copy only for these 10 URLs" rather than "generate Twitter cards for up to 1,000 credits." 2. Start with a small test batch. Run the agent on a handful of items first to confirm it behaves as expected before scaling up. 3. Review the task plan before confirming. Atlas Brain typically outlines what it intends to do before executing. Read this carefully and stop the task if it includes actions outside your intended scope. 4. Monitor credit usage during execution. Check your AI credit balance in the platform periodically during long-running tasks so you can catch unexpected consumption early. 🛠️ What to Do If Your Quota Has Been Consumed Unexpectedly If Atlas Brain has already consumed your AI credits in a way you did not intend, you can request a one-time quota reset from the Search Atlas support team. This is handled on a case-by-case basis by a specialised team. To request a reset: 1. Reach out to the Search Atlas support team directly. 2. Explain what task you ran, approximately how many credits you expected to use, and how many were actually consumed. 3. The team will review your account and, where appropriate, restore your quota so you can continue working.

Last updated on Oct 26, 2026

🛠️ Atlas Agent Error Affecting Playbook Functionality

🔍 Overview Some customers have experienced errors with Atlas Agent that prevent playbooks from running correctly. This article explains what caused the issue, what has been fixed, and what to do if you are still affected or would like to request compensation for the disruption. ⚙️ What Was Causing the Issue Several bugs affecting Atlas Agent were identified and have since been resolved by our engineering team. The confirmed issues included: - Atlas Agent failing in certain browsers that lacked full support for required browser features - Errors appearing when browsing all agent sessions or opening existing sessions - Unhelpful or raw error messages being surfaced to users instead of clear, readable ones - The tags field not functioning correctly when creating a new playbook All of the above have been marked as resolved. If you encountered any of these symptoms, the underlying cause has been fixed. ✅ Steps to Take If You Are Still Seeing Errors 1. Refresh your browser and clear your cache. Some error states can persist in the browser after a fix has been deployed. 2. Switch to a supported, up-to-date browser. One confirmed bug was specific to browsers missing certain modern browser capabilities. Using the latest version of Chrome, Edge, or Firefox is recommended. 3. Navigate back to Atlas Agent. Go to Left sidebar → Content → Content Genius and launch Atlas Agent from there to start a fresh session. 4. Retry your playbook. Open or create a playbook and test whether it runs successfully. If you were previously unable to use the tags field when creating a playbook, this should now be functional. 5. If errors persist, contact our support team. If you are still experiencing issues after following the steps above, our team needs to know so your case can be escalated. 💡 About Compensation for This Issue We understand that instability in a tool you rely on is frustrating. If Atlas Agent errors caused significant disruption to your workflow, you may be eligible for compensation. Our support team has the ability to escalate affected cases for review, and compensation has been offered to customers impacted by this bug. To be considered, you will need to contact us through the chat widget so your account can be reviewed and your case submitted for Tier 2 approval. Compensation decisions are made on a case-by-case basis. 🚀 What Our Team Did to Resolve This When this issue was first reported, our support team took the following actions: - Confirmed the bug was high priority and actively being worked on by developers - Logged affected customers so their cases were tracked - Escalated cases for compensation review through our internal process - Verified with customers once the fix was confirmed live If you reported this issue and have not yet heard back about compensation, please reach out through chat so we can follow up on your case. 📞 Still Need Help? If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

🚫 Cancel Wrongly Bulk-Published PRs and Recover HDC Credits

🔍 Overview If Atlas Agent triggered a bulk publication for the wrong press releases, act quickly. The publications can be stopped before they go live, and the HyperDrive Credits (HDC) spent on those placements can be refunded to your account. This article explains what happens, what you can do immediately, and how to get the issue fully resolved. ⚠️ Why This Happens Atlas Agent can initiate bulk PR publications rapidly. If the wrong press releases are selected before the action is confirmed, the publications may begin processing immediately. Once triggered, you cannot cancel them yourself from within the platform — intervention from the Search Atlas team is required to stop the placements and protect your credits. 🚀 Step 1 — Contact Support Immediately Speed matters. The sooner the team is notified, the more likely the publications can be halted before they go live. 1. Reach out to the Search Atlas support team as soon as possible. 2. In your message, include the following information: - Which press releases were published by mistake (titles, URLs, or any identifying details you have) - Approximately when the bulk publication was triggered - The number of HDC you believe were consumed Providing these details upfront allows the team to escalate to the PR team without delay. ⚙️ Step 2 — What the Team Does to Stop the Publications Once your request is escalated, the PR team will work on the back end to stop the affected press releases from being distributed to publishers. You do not need to take any action inside the platform for this step — the team handles it entirely on the back end. Once the team has acted, the affected PRs will no longer show as actively publishing. You can log in to the platform to check the status of your press releases. 💡 Step 3 — HDC Refund Processing After the publications are halted, the team will process a refund of the HDC consumed by the mistaken bulk publication. The refund is applied directly to your account balance — you do not need to submit a separate request for it. 📊 Step 4 — Verify Your Credit Balance Once the team confirms the refund has been processed, check your HDC balance inside the platform to confirm the credits have been restored. If the balance does not look correct after the team has confirmed the refund, reply to your support conversation and the team will investigate further.

Last updated on Oct 26, 2026

🔄 Fix Atlas Brain Local Citation Submission Loop

🔍 What Is Happening If Atlas Brain appears to hard-lock your HDC plan cap and repeatedly blocks local citation submissions — even though you are on a Pro plan with credits available — this is not a true hard lock. It is a sync issue between the Atlas Brain AI agent and the API. No credits are lost and no permanent restriction is applied to your account. ✅ Method 1: Start a Fresh Atlas Brain Chat A new Atlas Brain conversation resets the session state and clears the sync error. Follow these steps: 1. Close or exit your current Atlas Brain chat entirely — do not continue in the same thread. 2. Open a brand-new Atlas Brain chat session. 3. Re-initiate your local citation submission request from the beginning in the new session. Atlas Brain will re-read your account state with a clean call and should recognise your available credits, allowing the submission to proceed normally. 🛠️ Method 2: Submit Manually Through the Local SEO Tool If the fresh chat does not resolve the issue, you can bypass Atlas Brain entirely and submit your local citation directly through the platform's local SEO tool rather than through the AI agent. Navigate to the local citation submission area within the platform and complete your submission from there without using Atlas Brain. This manual route is always available as a reliable fallback regardless of any AI agent sync state. 💡 Why This Happens Atlas Brain can encounter a mismatch when reading quota or plan data during a session. This may cause it to incorrectly report a cap or refuse a submission it should allow. Starting a new chat or using the manual tool forces a fresh data read. This is a known intermittent behaviour and does not indicate a problem with your plan or credit balance. ⚙️ Before You Try Either Method — Quick Checks - Confirm you are on a plan that includes local citation submissions and that your credit balance is not genuinely zero. - Refresh the platform page before opening a new Atlas Brain chat, to ensure your browser session is also current. - If you have multiple projects, make sure you reference the correct site when restarting your request. 🙋 Still Blocked After Trying Both Methods? If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

🛠️ Project Missing From Atlas Agent Playbooks

🔍 Overview Some users have reported that a project they added to their account is visible in Site Explorer but does not appear in the Atlas Agent playbooks window, making it impossible to run playbooks for that project. In some cases, this same underlying sync issue can also cause additional symptoms such as backlink reports or other data tabs showing no data, or growth-plan features appearing locked. ⚙️ Is This a Known Issue? Yes. Missing projects in the Atlas Agent playbooks window is a known issue that the Search Atlas support team has investigated and resolved for affected users. The fix requires a backend correction applied by the support team to your specific account — it cannot be resolved by changing any setting on your end. ✅ Steps to Try First Before reaching out, try the following basic troubleshooting steps: 1. Hard-refresh your browser. Press Ctrl + Shift + R (Windows/Linux) or Cmd + Shift + R (Mac) to clear cached data and reload the latest version of the platform. 2. Try a different browser or incognito window. This rules out browser extensions or locally cached data as the cause of the missing project. 3. Log out and log back in. Sign out of your Search Atlas account completely, then sign back in. This ensures your session reflects the latest project data. 4. Verify the project is active. Confirm the project is still listed and active in your account. If it was recently added or modified, allow a few minutes for the platform to sync. 5. Open Atlas Agent and check the playbooks window again. Navigate to the Atlas Agent and look for your project in the playbooks window. 📋 What to Do If the Problem Persists If your project is still missing from the Atlas Agent playbooks window after completing the steps above, please contact our support team. A support agent will apply a backend fix to your account so the missing project appears correctly in the Atlas Agent playbooks window. This is the confirmed resolution for this issue and our team can typically get it sorted for you quickly. To help our team diagnose and fix the problem as fast as possible, please have the following information ready when you reach out: - The exact project URL or domain that is missing (for example, www.example.com). - Confirmation that the project is visible in Site Explorer. - A note about any additional symptoms you are experiencing, such as backlink reports showing no data or growth-plan features appearing locked. - A screen recording or screenshot showing the project is absent from the Atlas Agent playbooks window. To get in touch, open the chat widget in the bottom-right corner of the platform and type human teammate — we're happy to help.

Last updated on Oct 26, 2026

Fix AI Agent Pulling the Wrong Business URL

🔍 Overview If your AI Agent is referencing the wrong website URL or business details — for example, pulling a URL from a different domain than your active project — the issue is almost always in GBP Galactic, not in your project settings or OTTO connection. The AI Agent sources business information (including your website URL) directly from GBP Galactic, not from the Google Business Profile linked inside your OTTO project. This means that even if your project is correctly connected to the right GBP profile, the AI Agent will still use whichever GBP profile is activated inside GBP Galactic. ⚠️ Why This Happens You may have two or more GBP profiles connected across different domains in Search Atlas. A common scenario looks like this: - Project A is set to ericrounds.com with its own GBP profile. - Project B is set to a different domain (for example, a service business) with a separate GBP profile. - Both GBP profiles exist inside GBP Galactic, but the wrong profile is currently activated there. Because the AI Agent reads from GBP Galactic, it will use the active profile — even if that profile belongs to a different domain than your current project. Verifying or changing your project-to-GBP linkage in OTTO will not fix this. ✅ How to Correct the GBP Profile the AI Agent Uses To resolve this issue, you need to ensure the correct GBP profile is set as active inside GBP Galactic. Navigate to Local → GBP Galactic (GBP Projects) from the left sidebar in your Search Atlas account, review the GBP profiles connected there, and make sure the profile that corresponds to your active project's domain is the one currently active. If you are unsure which profile is active or how to switch it, please reach out to the Search Atlas support team and they can update the active GBP profile on your account to match your intended domain. 🧩 Understanding the Two Separate Connections It helps to think of these as two independent systems: - OTTO project → GBP connection: This controls which GBP profile is used for SEO tasks, optimizations, and project-level reporting inside OTTO. Changes here do not affect what the AI Agent sees. - GBP Galactic → Active profile: This is the data source the AI Agent reads from when generating responses, referencing your business URL, or personalizing content. This must be set correctly and independently. Both connections can point to different profiles without triggering an error, which is why the wrong URL can appear even when your project setup looks correct. 🛠️ Quick Checklist - You have more than one GBP profile connected in Search Atlas. - Your active OTTO project is set to the correct domain. - You have opened GBP Galactic and confirmed which profile is currently active there. - The active profile in GBP Galactic matches your intended project domain. If after reviewing GBP Galactic the wrong profile is still being used by the AI Agent, contact the Search Atlas support team so they can verify and correct the active profile on your account.

Last updated on Oct 26, 2026

🛠️ Agent Hub SEO Task Stalling or Stopping Early

🔍 Overview Some customers have reported that the Fix Critical On-Page SEO Issues task in Agent Hub runs for approximately an hour, stops after completing only the first step, and generates an unusually large amount of text output. If this has happened to you, the most likely cause is a temporary outage with one of the AI providers that powers the platform — not an error in your project or settings. ⚙️ Why This Happens Agent Hub SEO tasks rely on third-party AI providers to process and complete each step in the workflow. When one of those providers experiences instability or an outage, the task can stall mid-run, time out, or behave unexpectedly — including stopping after the first step and producing excessive text output. These outages are temporary and are resolved by our engineering team as quickly as possible. Our team has confirmed that this type of stalling behaviour is a known issue during AI provider outages and has been addressed in previous engineering updates to improve how the agent handles multi-step workflows. ✅ What To Do 1. Wait a short time before retrying. AI provider outages are usually temporary. Waiting 30–60 minutes before trying again gives providers time to recover. 2. Re-run the task from scratch. Once you believe enough time has passed, navigate back to Agent Hub and start the Fix Critical On-Page SEO Issues task again. Do not attempt to resume a stalled run — start a new one. 3. Monitor the task closely on your next attempt. If the task completes all steps successfully this time, the provider issue has been resolved. If it stalls again in the same way, the outage may still be ongoing. 4. Check whether the issue persists across multiple attempts. If the task consistently fails over several hours or the next day, it is no longer likely to be a temporary outage and you should contact support. 🚩 Signs That This Is an AI Provider Issue - The task runs for a long time (around an hour) before stopping. - Only the first step completes; subsequent steps do not run. - The output contains an unusually large or repetitive amount of text. - The same task has worked correctly for you in the past. - The issue resolves itself after waiting and retrying. 💡 Tips to Avoid Confusion - Do not run the task multiple times simultaneously. Starting duplicate runs during a provider issue will not help and may compound the problem. - Do not change your project settings in response to this error — the stalling is not caused by your configuration. - Give it at least a day before escalating if retries continue to fail, as some provider outages can last several hours. 📞 Still Need Help? If you have waited, retried, and the task is still stalling or stopping early, our team is ready to investigate further. If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

🛠️ Fixing the 400 Error in Agent Hub Custom Prompts

🔍 What This Error Means A 400 error in Agent Hub means the request your custom prompt sent could not be processed by the system. It is not a problem with your account or with Agent Hub itself. Instead, it usually points to something in the way the custom prompt is written or configured. This is why playbooks keep working while your custom prompt fails. Playbooks are pre-built and validated by Search Atlas, so their structure, variables, and formatting are already correct. A custom prompt is created from scratch, which means a small formatting or input issue can cause the request to be rejected. ⚙️ Common Causes of the 400 Error Most 400 errors with custom prompts come from one of the following: - Prompt is too long: Very long prompts can exceed the allowed input size and be rejected. Prompts that go over the character limit are a confirmed cause of this error, especially when they contain long or detailed instructions. - Unsupported characters: Pasting from another program can add hidden symbols, smart quotes, or formatting that the system cannot read. - Empty or missing fields: A required field, such as a target URL or keyword, was left blank. - Broken variables: A placeholder or variable was edited incorrectly, so the system cannot fill it in. - Invalid links: A URL in the prompt is misspelled, incomplete, or missing the https:// prefix. 🛠️ How to Fix It Work through these steps in order. Most customers resolve the error within the first few checks. 1. Shorten your prompt: Remove any text that is not essential and keep your prompt within the character limit. Keep instructions clear and direct. 2. Retype instead of pasting: If you copied the prompt from a document or email, retype it directly into Agent Hub to strip out hidden formatting. 3. Check all required fields: Make sure every field the prompt needs is filled in correctly and nothing is left blank. 4. Review your variables: Confirm that any placeholders match the format used in working playbooks. Do not change the surrounding symbols. 5. Verify your links: Make sure each URL is complete, correctly spelled, and begins with https://. 6. Test with a simple prompt: Enter a short, basic prompt to confirm the error is tied to your original wording and not a wider issue. 💡 Tips to Prevent Future Errors A few habits will help your custom prompts run smoothly every time: - Start from a playbook: Use a working playbook as a template, then adjust it gradually to fit your needs. - Build in small steps: Add one instruction at a time and test as you go, so you can quickly spot what causes a problem. - Keep prompts focused: Clear, concise prompts are less likely to trigger errors and produce better results. - Avoid special formatting: Stick to plain text and avoid pasting styled content from other tools. 💬 Still Seeing the Error? If you have tried these steps and the 400 error continues, our team is ready to help. If you need further assistance, click the chat icon in the bottom-right corner of the platform to start a live chat with a member of our team. Sharing the exact prompt you used will help us resolve the issue faster.

Last updated on Oct 26, 2026

🔌 Troubleshooting AI Agent Connection Errors on Send

This article helps you understand what to do if you encounter a connection error when submitting a prompt in an AI-powered area of the platform. Because the specific steps that resolve this error may vary depending on your account and environment, the best path forward is to reach out directly so our team can investigate and fix it for you. 📍 Where This Occurs The connection error typically surfaces after you submit a prompt in an AI-powered content area of the platform. The same error is returned regardless of what you type, which indicates a session or agent connectivity issue rather than a problem with your prompt itself. 🛠️ What to Have Ready When You Contact Support To help our team resolve this as quickly as possible, please have the following information ready before reaching out: - The name of the project or workspace you were working in when the error occurred - The exact error message you see on screen (copy the full text if possible) - The date and approximate time when the error first appeared - The browser and operating system you are using - Whether the error appears consistently on every prompt submission, or only sometimes ✅ What Happens Next Once you contact our team with the details above, a member of our support team will investigate the connection issue on the backend and work to restore normal function for your account. If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team. Additional Notes Any recent changes made to your account, integrations, or agent configuration

Last updated on Oct 26, 2026

Missing Skill in Your Migration Workflow

If a skill component appears to be missing while working through a multi-stage migration process in Search Atlas, use this article to collect the right information and get targeted help from our support team. ⚠️ Why This Happens Migration workflows may involve packaged skill files or templates that are applied across multiple stages. A missing component can occur for a number of reasons — including an incomplete download, a mismatch between your account configuration and the expected package version, or a feature that has not yet been enabled for your workspace. Because the root cause varies by account, there is no single universal fix, and our team will need to review your specific setup. The most commonly reported missing component is the migration-template-builder skill, which is expected at Stage 4 of the workflow. The v1.4 package (released 2026-06-23) has been confirmed in some cases to ship with only six skill folders, leaving this component absent. The underlying feature has been completed on the engineering side, so if this skill is missing from your package, the most likely cause is that it has not yet been enabled for your specific workspace — our support team can arrange this directly. 📋 Information to Collect Before Contacting Support Having the following details ready will help our team resolve your issue faster: - The name of the missing skill or template component, exactly as it appears (or should appear) in the workflow — for example, migration-template-builder at Stage 4 - The stage or step in the migration process where the issue occurred - Whether you received or downloaded a package or file as part of this workflow, and when — including the version number if visible (for example, v1.4) - A screenshot showing the missing element or any error message displayed on screen - Your account email address and the name of the affected workspace or project 🔄 Basic Troubleshooting Steps Before reaching out, try the following to rule out common session or cache issues: 1. Refresh your browser and re-open the migration workflow from the beginning. 2. Log out of your Search Atlas account completely, then log back in. 3. Try accessing the workflow in a different browser or in an Incognito / Private window. 4. If you downloaded a package or template file, confirm the download completed without errors. If uncertain, attempt to download it again from the original source. 5. If the skill is still missing after these steps, contact our live support team with the details gathered above. ⏱️ What to Expect From Support Our team will review your account and migration workflow configuration to identify why the skill component is not available. For issues that require a deeper investigation, your case will be escalated to the relevant product team. You will be kept updated throughout the process and notified once a resolution or workaround is confirmed. 💬 Need More Help? If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

🛠️ Fix Atlas Brain and AI Quota Errors

🔍 Overview If you see a 'No available quota for AI Visibility Analysis' error or find that Atlas Brain is unresponsive, the cause is almost always a mismatch between your visible AI quota and the features actually included in your current plan. This article explains why this happens and how to fix it. ⚙️ Why This Error Occurs Your account dashboard may show an AI quota balance (for example, 500 credits), yet Atlas Brain still refuses to run. This contradiction usually happens for one of the following reasons: - Plan downgrade or change: Your subscription was changed to a plan that no longer includes the Atlas Brain feature or LLM Visibility Analysis, even though leftover quota credits are still displayed. - Feature not included in current plan: AI quota credits and Atlas Brain access are separate entitlements. Having quota credits does not automatically unlock every AI-powered feature — Atlas Brain must be explicitly included in your plan. - Billing cycle reset: A recent renewal or plan modification may have reset feature access before your quota was fully refreshed, creating a temporary mismatch. - Account configuration error: In rare cases, a backend sync issue can incorrectly flag your account as ineligible for a feature you should have access to. 🚀 How to Troubleshoot Step by Step 1. Check your current plan: Go to Avatar (top-right) → Billing → Plans & Top-ups and confirm which plan you are on. Look specifically for whether Atlas Brain or LLM Visibility Analysis is listed as an included feature. 2. Review your AI quota details: On the same page, expand your quota breakdown. Confirm that the quota shown is allocated to the feature you are trying to use — not just a general credit pool. 3. Log out and log back in: A fresh session forces your account entitlements to resync. After logging back in, retry Atlas Brain before taking further steps. 4. Clear your browser cache: Stale cache can cause the dashboard to display outdated quota or feature data. Clear your cache, reload the platform, and try again. 5. Try a different browser or incognito mode: This rules out browser extensions or local storage issues interfering with feature access. 6. Confirm no recent plan changes were made: If someone else manages billing on your account, verify no downgrade or seat change occurred recently that may have removed Atlas Brain access. 💡 Understanding AI Quota vs. Feature Access It is important to understand the difference between AI quota credits and feature access: - AI quota credits are the numerical allowance consumed each time an AI-powered action runs (such as generating content or running an analysis). - Feature access is a plan-level permission that determines which tools and modules are available to your account at all. Think of it this way: quota credits are the fuel, but feature access is the key that starts the engine. If Atlas Brain is not included in your plan, having credits will not unlock it. You need both to use the feature successfully. 🔄 How to Restore Atlas Brain Access If you confirm that Atlas Brain or LLM Visibility Analysis is missing from your current plan, here are your options: - Upgrade your plan: Navigate to Avatar (top-right) → Billing → Plans & Top-ups and select a plan that includes Atlas Brain. Changes take effect immediately after checkout. - Contact our team to review your account: If you believe Atlas Brain should already be included in your plan based on your purchase, our team can investigate and correct any configuration errors on the backend. 🛡️ Preventing This Issue in the Future - Before changing or downgrading your plan, review the feature comparison table to confirm Atlas Brain remains included. - If multiple team members manage billing, communicate plan changes in advance to avoid unexpected feature loss. - Bookmark the Avatar (top-right) → Billing → Plans & Top-ups page so you can quickly verify your entitlements whenever an unexpected error appears. 🤝 Still Need Help? If you have followed all the steps above and Atlas Brain is still not working, or if your plan appears correct but the error persists, our team is ready to investigate your specific account. If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

🛠️ Classic View Toggle Unclickable in Agent Hub

🔍 Overview Some users found that the Classic view toggle in the left sidebar under Home → Classic could not be clicked, making it impossible to switch back to Classic mode. This was caused by a full-screen loading overlay that appeared while playbooks were loading, blocking interaction with the toggle button. This issue has been identified, tracked, and resolved by the Search Atlas engineering team. If you encountered this problem before, the fix is now live and you should be able to use the toggle normally. ⚠️ What Was Happening When the Agent Hub homepage loaded, a full-screen playbook loading overlay would sometimes appear and remain on top of the page. This overlay blocked click interactions with the Classic view toggle, even though the button was visible on screen. The result was a toggle that looked active but could not be selected. Additional related issues that were fixed at the same time include: - Non-clickable areas on the Classic and Agent Hub toggle button on the homepage - Classic mode toggle not appearing at all for some Classic View users - Homepage connector and scheduled job rows being unresponsive to clicks ✅ What To Do Now 1. Refresh your browser to make sure you are running the latest version of the platform. 2. Navigate to the Agent Hub homepage. 3. Try clicking the Classic view toggle again to switch between views. 4. If the toggle responds and switches the view as expected, the fix is working correctly for your account. 🔄 If the Toggle Is Still Unresponsive In rare cases, a cached version of the page may still cause the issue to appear. Try the following before contacting support: - Hard refresh your browser (Ctrl + Shift + R on Windows, Cmd + Shift + R on Mac). - Clear your browser cache and cookies, then reload the platform. - Try a different browser or an incognito/private window to rule out a browser-specific conflict. - Check whether the issue occurs on a different device or network. If none of these steps resolve the problem, the issue may be account-specific and will need to be reviewed by our team. 💬 Need Further Help? If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

🔍 Fix a Missing Project in Agent Hub

🧭 Overview If you have projects in Search Atlas but one or more do not appear in the Agent Hub project list, there are several self-serve steps you can try to resolve the issue before reaching out to support. 🔎 Why a Project May Not Appear A project missing from the Agent Hub project list is typically not related to account permissions or access restrictions. If the project is visible in your main Search Atlas dashboard, your account access is intact. The root cause is generally something within the Agent Hub setup for that specific project, but the precise reason varies by case. ✅ Self-Serve Steps to Try First Before contacting support, work through the following steps in order: 1. Refresh the Agent Hub page. Navigate away from Agent Hub and return to it, or do a hard refresh (Ctrl+Shift+R on Windows, Cmd+Shift+R on Mac) to rule out a browser display issue. 2. Check the project in your main Search Atlas dashboard. Confirm the project is active and fully set up there. If it was recently created, allow a few minutes for it to propagate to Agent Hub before checking again. 3. Verify your project count. Count the total number of projects visible in your Search Atlas dashboard and compare that to what appears in Agent Hub. Note the discrepancy so you can report it accurately if escalation is needed. 4. Log out and log back in. Sign out of Search Atlas entirely, clear your browser cache, then sign back in and recheck Agent Hub. 5. Try a different browser or incognito window. This helps rule out browser extension or cache conflicts affecting what projects load in Agent Hub. If the project still does not appear after completing all of the steps above, the issue likely requires a backend investigation by our support team. 📋 What to Have Ready When You Contact Support To help our team resolve this as quickly as possible, please have the following information ready before reaching out: - The exact name of the missing project as it appears in your Search Atlas dashboard. - The total number of projects in your Search Atlas account and how many are currently visible in Agent Hub. - Approximately when you first noticed the project was missing. - Whether the project is newly created or was previously visible in Agent Hub and has since disappeared. - Any error messages you may have seen, with the exact wording. - Confirmation of which self-serve steps above you have already completed. 🚫 What This Issue Is Not A missing project in Agent Hub is not caused by deactivated accounts, incorrect permissions, or access restrictions. If you can see the project in the main Search Atlas dashboard, your account access is fine. If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

⚡ AI Recommendation Generation: Batch Size Limits & Timeout Fixes

Overview When you run AI recommendation generation across a large set of pages, jobs that include too many pages in a single batch may time out before completing. This article explains the cause and the recommended approach for working through large page sets without losing progress. Why Large Batches Time Out AI recommendation generation sends each page through a processing pipeline that analyses content, evaluates on-page signals, and produces actionable suggestions. Each page adds to the total processing time for the job. When a single batch contains too many pages, the cumulative processing time can exceed the system's job timeout threshold, and the job may be cancelled before all pages are completed. - Pages already processed before a timeout are not lost — their recommendations are saved and visible once the job ends. - Remaining unprocessed pages will show no recommendations until you rerun generation for them. Recommended Batch Size To ensure every job completes successfully, limit each AI generation run to a smaller subset of pages rather than selecting your entire page list at once. Smaller batches process well within the timeout window and give you predictable, reliable results. You can run multiple batches back-to-back — each time selecting only pages that have not yet been processed. How to Process Large Page Sets in Batches Follow these general steps to work through a large number of pages without hitting the timeout limit. 1. Open the AI generation tool you are using. Navigate to the relevant tool in your Search Atlas account where you run AI recommendation generation. 2. Filter for pages without recommendations. Use any available filters to display only pages that have not yet received AI-generated recommendations. This keeps your working list clean and prevents duplicate processing. 3. Select a small subset of pages. Choose a manageable batch of pages from your filtered list. Keeping batches small reduces the risk of hitting the timeout threshold. 4. Trigger AI generation for your selection. Start the generation job for your selected pages and allow it to run to completion before starting the next batch. 5. Wait for the job to finish. Once the batch is complete, the recommendations for those pages will be available. 6. Repeat for the next batch. Return to your filtered list, select the next set of unprocessed pages, and repeat the process until all pages have been covered. What to Have Ready if You Need Support If you continue to experience timeouts even with smaller batches, please have the following information ready when you contact support: - The name of the project and the specific tool where the issue is occurring - The approximate number of pages you were processing when the timeout occurred - Any error messages or job status indicators you observed If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

AI Agent Loop Drained Your Article Quota

🔍 What Happened In some cases, the AI agent inside Content Genius can enter a loop where it repeatedly creates and recreates articles without stopping. Each cycle consumes your article quota and premium AI quota, even though no usable content is produced. This is a known platform issue that our engineering team is actively working to resolve. ⚠️ Can You Fix This Yourself Quota restoration cannot be done from within your account settings — it requires a backend credit applied directly by our support team. However, you can take two immediate steps right now to stop further quota loss and speed up your reset: 1. Stop the AI agent immediately. As soon as you notice the loop, stop or close any active AI agent task that is running to prevent additional quota from being consumed while you wait for the reset. 2. Document what you see. Take a screenshot of your current article list or quota usage screen and note the approximate time the loop started and the project or site affected. This information will allow our team to process your reset faster. Once you have stopped the agent and gathered those details, use the chat widget at the bottom of this page to message our support team directly. Paste your screenshot, the date and time the loop started, and the affected project name into your first message. A team member will review your account activity, confirm the quota consumed by the loop, and apply the backend credit — you will be notified as soon as the reset is complete so you can resume working right away. If you do not receive a response within a few minutes during business hours, type human teammate in the chat widget to be connected with a member of our team. 🛠️ What Our Team Will Do When you contact support, a team member will: 1. Review your account activity to confirm the loop occurred and identify how much quota was consumed. 2. Perform a backend reset of your affected content and premium AI quotas. 3. Confirm once the reset is complete so you can resume working. This process is straightforward and our team can action it quickly once they have reviewed your account. 📋 Before You Reach Out To help our team resolve this as fast as possible, have the following ready when you open chat: - Approximate date and time the loop occurred. - The project or site the AI agent was running on. - Any article titles or keywords the agent was working with, if you remember them. - A brief description of what you observed — for example, articles being created multiple times in a row. 💡 How to Avoid This in the Future While our engineering team works on a permanent fix, you can reduce the risk of a loop consuming your quota by: - Monitoring AI agent tasks closely when first starting a large batch of article generation. - Stopping the agent manually if you notice the same article being created more than once. - Running smaller batches rather than large site-wide generation runs so that any loop is caught earlier with less quota impact.

Last updated on Oct 26, 2026

Website Studio AI Agent Creating Rogue Pages

🔍 What Is This Issue? In some cases, the Website Studio AI agent may behave unexpectedly — publishing content automatically, generating repeated pages you did not request, or appearing to loop in a way you cannot stop. Pages created during this state may persist even after you attempt to remove them. This is a known issue that has been reported by other users and has been escalated to the engineering team for investigation. ✅ What To Do If You Are Experiencing This If you see unexpected pages, repeated entries, or content you did not create in Website Studio, follow these steps to stop the behaviour and help the team resolve it: 1. Stop using the AI agent immediately. Do not submit new prompts while the issue is active. Additional instructions may trigger further unintended page creation. 2. Document what you see. Take screenshots of the unexpected pages, including their titles and URLs if visible. This will help the support team investigate and clean up the affected records on the backend. 3. Do not manually delete the rogue pages yet. Manual removal on your end may not fully resolve the issue, as the underlying records may need to be cleared by the support team. Wait for confirmation before removing anything. 4. Identify the affected pages. Review your Website Studio project and confirm which pages were created unexpectedly. Note the exact page names and the approximate time they were generated — the support team will need this information to locate and remove the affected backend records for your specific project. 5. Contact our support team with the screenshots, page names, and timestamps you collected. Our team can access your project on the backend, remove the rogue page records, and confirm when it is safe for you to continue using the AI agent. If you don't hear back promptly, you can also reach us by replying to any previous support email or starting a new conversation through the support messenger. 💡 Tips To Avoid Similar Issues - Use clear, specific prompts. Vague or overly complex instructions can cause the AI agent to misinterpret your intent. Describe your goal as precisely as possible — for example, specify whether you want to edit an existing section or create a new page. - Review AI actions before confirming. After the agent proposes changes, review the summary carefully before approving, especially if it references page creation when you only intended to edit a section. - Work on one task at a time. Avoid sending multiple layered instructions in a single prompt. Break complex tasks into smaller, sequential steps. 🆘 Need Further Help? If you need further assistance, please don't hesitate to reach out to our support team and we will be happy to help.

Last updated on Oct 26, 2026

🛠️ Troubleshooting Atlas Agent 400 Error on Hero Image Prompts

This article addresses the request failed: 400 error that can appear in Atlas Agent when submitting a hero image prompt. A 400 error means the request was rejected before it was processed. Before reaching out to our team, try the self-serve steps below — many users are able to resolve this by adjusting their prompt. 🔧 Self-Serve Steps to Try First 1. Retry the prompt: Occasionally a transient issue can cause a 400 error. Submit the exact same prompt again to see if it succeeds on a second attempt. 2. Simplify the prompt: Shorten your hero image prompt to the core description, removing any extra instructions, punctuation, or formatting. A very long or complex prompt can sometimes trigger a rejection. 3. Check for unsupported characters: Remove any special characters (e.g., /, &, <, >, quotes within quotes) from the prompt text and try again. 4. Try a different prompt entirely: Write a new, plain-language description of the hero image you need. If this succeeds, the issue is specific to something in your original prompt wording. 5. Reload the page and try again: Refresh your browser, navigate back to the Atlas Agent project, and resubmit the prompt to rule out a session-level issue. 📋 If the Error Persists — What to Prepare Before Reaching Out If none of the steps above resolve the error, our team will need to investigate on a case-by-case basis. Please have the following ready: - Your project name: The exact name of the project where the error occurred. - The exact error message: Copy the full error text as it appears on screen (e.g., request failed: 400). - The hero image prompt you submitted: Include the full text of the prompt that triggered the error. - Timestamp: Note the date and approximate time when the error occurred. - Steps you already tried: Any actions you took before or after the error appeared, including which self-serve steps above you attempted. ✅ Next Steps Having the information above ready will allow our team to identify the cause and apply the correct fix without unnecessary back-and-forth. If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

🛠️ Fix the Outreach Campaign Schedule Error

🔍 What This Error Means When you see the message 'Please configure outreach schedule first!', it means your outreach campaign is missing a required sending schedule. Search Atlas will not allow a campaign to launch until a valid schedule is set. This is a quick fix — follow the steps below to resolve it. 🛠️ How to Configure Your Outreach Schedule 1. In the left sidebar, open your outreach campaigns list. 2. Locate the campaign showing the error and click on it to open the campaign settings. 3. Navigate to the Schedule section within the campaign editor. 4. Set your preferred sending days, time window, and time zone. All three fields are required before you can proceed. 5. Click Save to confirm your schedule settings. 6. Return to the campaign overview and attempt to launch your campaign again. ✅ Tips for a Valid Schedule - Select at least one sending day — campaigns with no active days cannot be scheduled. - Set a valid time window — your start time must come before your end time. - Choose a time zone — leaving the time zone blank will trigger the same error even if other fields are filled in. - Save before launching — always save your schedule changes before clicking the launch button. 🚀 Launching Your Campaign After Setup Once your schedule is saved, the error should no longer appear. Go back to your campaign in your campaigns list, confirm the schedule is displayed correctly in the campaign summary, then click Launch. Your campaign will begin sending according to the schedule you configured. 💡 Still Seeing the Error? If you have completed all the steps above and the error persists, it is possible that the schedule did not save correctly due to a browser or session issue. Try the following: - Refresh the page and re-enter your schedule settings before saving again. - Clear your browser cache or switch to a different browser. - Check that no required fields are highlighted in red, which indicates an incomplete entry. If the issue continues after trying these steps, If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

🔧 Fix Coworker Slack Visibility and Switch Credits

Overview Two common questions arise together: Coworker not appearing inside the Search Atlas Slack integration, and how to switch from legacy credits to Universal Credits. This article explains what is happening, what to have ready when you contact support, and what to expect during your credits migration. Why Coworker Is Not Visible in Slack If Coworker is not appearing as expected within your Slack integration, our team investigates these cases on a per-account basis and will need to take action on the backend to resolve it. When contacting support about this issue, please have the following ready: - Your Search Atlas account email or workspace name. - The exact behavior you are seeing (e.g. Coworker not appearing, unable to connect, an error message — include the exact wording if possible). - The timestamp of when you first noticed the issue. Universal Credits Migration If you need to switch from legacy credits to Universal Credits or have questions about your credits balance after a migration, our team can assist with the transition on the backend. When contacting support about credits migration, please have the following ready: - Your Search Atlas account email or workspace name. - Whether you currently see a legacy credits balance and approximately how many credits remain. - Any error message or unexpected behavior you encountered during the migration process, including exact wording and timestamp. If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

💬 How to Locate Missing Chat History

This article explains what to do when a previous chat session is not visible in the platform. If you believe you did not delete a conversation but it is no longer appearing, the steps below will help you locate it and escalate if needed. Common Causes of Missing Chat History - Account or workspace mismatch: You may be logged into a different account or workspace than the one where the original chat took place. - Browser or cache issue: A corrupted browser cache can sometimes prevent chat history from displaying correctly. Steps to Try 1. Confirm you are signed into the correct account. If you use multiple workspaces or email addresses, log out and sign back in with the account that was active during the original chat. 2. Perform a hard refresh of the page (press Ctrl + Shift + R on Windows or Cmd + Shift + R on Mac) to clear any cached display issues. 3. Clear your browser cache and cookies, then reload the platform and check whether the chat history reappears. 4. Try accessing the platform in a different browser or an incognito/private window to rule out a local browser issue. If the Chat History Is Still Missing If the conversation does not reappear after completing the steps above, please reach out to our support team with the following information ready so we can investigate on our end: - The approximate date and time of the missing conversation - The account email address or workspace name associated with the session - A description of where in the platform you expected to find the chat history If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

⚡ AI Agent Infrastructure Cleanup and Code Improvement

As part of a recent AI Agent release, an internal code cleanup improvement was shipped. Unused legacy components were removed from the codebase as they were no longer serving any active function in the platform. What's New Internal legacy components and associated dead code have been removed. These components were no longer called or used anywhere in the platform. Removing them reduces unnecessary overhead and helps keep the codebase clean and stable. What This Means for You No action is required on your part. This is a behind-the-scenes code cleanup that takes effect automatically. Your experience using the AI Agent remains unchanged — no features have been removed and no workflows are affected. Noticing Something Unexpected After This Update? While this cleanup has no customer-facing impact, if you believe you are seeing unexpected AI Agent behavior around the time of this release, it is most likely unrelated to this change. To help our team investigate efficiently, please have the following ready before reaching out: - Your project name - A clear description of the unexpected behavior you are seeing - The approximate time the issue first occurred Our team can use these details to look into your specific account and determine the cause quickly. If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

🔌 Atlas AI Agent 400 Error – Search Console Fix

🔍 What You May Be Experiencing If you recently noticed your website traffic data drop suddenly — for example, from 1,000–2,000 daily visits down to around 40 — and you are also seeing a 400 error when running prompts in the Atlas AI Agent, these two issues share the same root cause: a stale or broken Google Search Console (GSC) integration. When the GSC connection becomes disconnected or stops syncing correctly, the Atlas AI Agent loses access to your live traffic and performance data. This causes prompt failures and makes your traffic appear to have collapsed inside the platform. ⚠️ Common Symptoms - Atlas AI Agent returns a 400 error when you submit a prompt - Daily website traffic shown in the platform drops dramatically and unexpectedly - Site Metrics data appears outdated, incomplete, or frozen - AI Visibility or performance insights seem inaccurate or missing 🧩 Why This Happens Google Search Console connections can become stale over time due to expired OAuth tokens, permission changes in your Google account, or property mismatches. When this happens, Search Atlas can no longer pull fresh data from GSC. The Atlas AI Agent relies on this data to respond to prompts — and without a valid connection, it returns a 400 error instead of a useful answer. This is not a problem with your website itself. Your actual traffic has not dropped. The platform simply cannot display or use your data while the GSC connection is broken. ✅ How to Fix It – Reconnect Google Search Console Follow these steps to disconnect and reconnect your Google Search Console integration: 1. Click your avatar in the top-right corner of the platform to open the Avatar Menu (👤). 2. Navigate to Settings → Integrations — Note: if your GSC integration settings are housed under a dedicated Integrations or Connected Accounts section visible in your account, navigate there instead. 3. Locate your Google Search Console connection and click Disconnect. 4. Once disconnected, click Connect (or Reconnect) and follow the Google OAuth prompts to sign in and grant permissions. 5. When selecting a property during reconnection, make sure you choose the exact GSC property that matches your website — for example, https://www.yourdomain.com vs. http://yourdomain.com. A property mismatch is a common cause of sync failures. 6. Save your settings and return to Site Metrics (left sidebar → Site Metrics) to confirm your data is loading correctly. ⏳ Data Sync Timeline After reconnecting Google Search Console, please allow 24 to 48 hours for your traffic and performance data to fully resync inside Search Atlas. During this window: - Traffic numbers may still appear lower than expected — this is normal while the sync catches up. - The Atlas AI Agent should resume responding to prompts correctly once the connection is validated, though data-dependent answers will improve as the sync completes. - If you still see a 400 error after 48 hours, proceed to the troubleshooting step below. 🔎 Verify Your GSC Property Is Correct A common reason reconnection does not resolve the issue is a property mismatch. In Google Search Console, the same domain can exist as multiple properties (e.g., with or without www, with http or https). Make sure the property you connect in Search Atlas is: - The property that is verified in your Google Search Console account - The property that has active data — check inside GSC directly to confirm traffic is visible there - The primary version of your domain (the one your visitors actually land on) 🤖 About Atlas AI Agent and AI Visibility The Atlas AI Agent is your in-platform AI assistant that answers questions about your site's SEO performance, traffic trends, keyword rankings, and more. It pulls data in real time from connected sources — including Google Search Console — to generate accurate, personalised answers. AI Visibility is a feature within Search Atlas that tracks how your brand and content appear across AI-generated search results and answer engines. Both features depend on a healthy GSC connection to surface accurate insights. Reconnecting GSC restores full functionality for both. 🛠️ Still Seeing a 400 Error? If you have completed the reconnection steps and waited 48 hours but the 400 error persists when using the Atlas AI Agent, there may be a permissions issue with your Google account or a configuration problem that requires a manual review. If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

🔍 Troubleshooting Atlas Agent Playbooks

This guide helps when Atlas Agent opens but its playbooks are missing or inaccessible. What This Issue Looks Like When accessing Atlas Agent, you may find that playbooks are not listed, fail to load, or return an access error. This prevents you from using the playbooks that should be available in your workspace. Steps to Try First Before contacting support, work through the following steps. Missing or inaccessible playbooks are often caused by being logged into the wrong workspace or a stale session, which you can resolve on your own: 1. Confirm which workspace you are in. Check the workspace name shown in the platform and compare it to the workspace where your playbooks were created. If you belong to more than one workspace, switch to the correct one and reopen Atlas Agent. 2. Verify your access and plan. Confirm that your plan includes Atlas Agent, and ask a workspace owner or admin to verify that your role has permission to view playbooks. If your access was recently changed or removed, they can restore it for your account. 3. Refresh your session. Log out of the platform and log back in, or open Atlas Agent in an incognito or private browser window. This rules out a cached session preventing playbooks from appearing. 4. Verify that playbooks load. Reopen Atlas Agent and confirm that the expected playbooks are listed and open without an access or loading error. What to Have Ready When You Contact Support If the steps above do not restore access, the issue may require investigation on the backend. Please have the following information ready before reaching out so the issue can be resolved as quickly as possible: - Your workspace name - A description of what happens when you open Atlas Agent (e.g. playbooks missing, error shown, page blank) - The exact error message, if one appears - The date and time when the issue first occurred - A screenshot showing the issue, if available - Which of the steps above you already tried and what happened What to Expect Once the support team reviews your workspace and the affected playbooks, they will take the necessary steps to restore access. You should then be able to open Atlas Agent and see the expected playbooks listed without any access or loading errors. If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

🛠️ Fix Atlas Brain Chart and Diagram Rendering Issues

What Is This Issue? When Atlas Brain generates a document that includes a chart, diagram, or table, the output can sometimes appear as broken symbols, raw markup, or unreadable text instead of a clean visual. This is a known rendering issue reported by users — the underlying data is correct, but it is not displaying in the expected format. Common signs you are experiencing this issue: - A chart or diagram appears as a wall of symbols, brackets, or encoded characters. - A table loses its structure and runs together as unformatted text. - The document appears correct at first but breaks when you scroll, copy, or revisit it. What to Do Because the root cause of this issue requires investigation by our team, there are no guaranteed self-serve steps to resolve it. However, the following actions may help in the short term and will also gather the information our support team needs: 1. Note the exact document or session where the issue occurred. If you have a session link or document name, copy it — our team will need this to investigate. 2. Refresh the page. Reload the platform and reopen the document to see if the rendering corrects itself. 3. Try a different browser. If the garbled output persists, open Search Atlas in a different browser to rule out a local browser compatibility issue. 4. Ask Atlas Brain to reformat the output. In the same session, you can try prompting the agent to present the same information in a different format, such as a plain list or a descriptive paragraph, as a temporary workaround while the underlying issue is investigated. 5. Capture a screenshot. Take a screenshot of the garbled output so our team can see exactly what you are experiencing. When to Escalate If the issue persists across browsers and sessions, please reach out to our support team. To help us investigate quickly, have the following ready: - The session link or document name where the issue occurred. - A screenshot of the garbled output. - A description of what you asked Atlas Brain to generate (for example, the prompt you used). - The approximate date and time the issue first appeared. If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be instantly connected with a member of our team.

Last updated on Oct 26, 2026

🔧 Troubleshooting AI Schema Generation Delay False Negatives

This article addresses a situation where AI-generated page-level schemas appear to complete successfully — credits are deducted and the platform confirms generation — but no schema output is visible when you check your pages. If you are experiencing this, the steps below will help you gather the information needed to resolve it. 🔍 Common Causes - Processing lag: Large batch jobs can take several minutes to fully propagate. The platform may confirm the job before all outputs are written. - UI cache: The page you are viewing may be showing a cached state from before the generation completed. - Partial job failure: In rare cases, the generation job starts and deducts credits but encounters an error mid-batch, leaving some or all pages without output. 🛠️ Step-by-Step 1. Wait a few minutes before taking further action. A delay between the confirmation message and results appearing in the interface is possible. Do not retry immediately. 2. Hard-refresh the page. Press Ctrl + Shift + R (Windows/Linux) or Cmd + Shift + R (Mac) to bypass the browser cache and reload the current view with fresh data. 3. Navigate away and return. Open a different section of the platform, then navigate back to the schema results page. This prompts the interface to load the latest data. 4. Check whether schema outputs have appeared. Return to the page or section where you submitted the generation request and look for any schema entries associated with the pages you submitted. 5. If outputs are still missing, do not retry the batch yet — a duplicate run may deduct additional credits. Instead, note the exact time of generation and the number of pages submitted, and contact our support team with that information. ✅ What to Have Ready When You Contact Support - The exact date and time you submitted the generation job. - The number of pages included in the batch. - A description of what you see in the platform (e.g., confirmation message received, but no schema output visible). - Your current credit balance, so our team can verify whether a credit adjustment is needed. If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

🕸️ Domain Knowledge Network Empty Despite Complete Setup

🔍 Overview If your Domain Knowledge Network (DKN) is showing no data even though your knowledge graph completion is at 100% and all required business information fields are filled in, you are not alone. This is a known issue that can occur after setup is complete. The problem is not caused by missing information — it appears to be related to a generation or display issue on the platform side that requires investigation by our support team. ✅ Confirm Your Setup Is Complete Before reaching out, quickly verify the following are in place: - Knowledge graph completion shows 100% in your dashboard. - All required business information fields are filled, including business name, website URL, business category, target location, and a business description. - Your project has been saved and there are no unsaved changes pending. If all of the above are confirmed and your DKN is still empty, this is something our support team will need to investigate directly on your account. 📋 What to Have Ready When You Contact Support To help our team resolve this as quickly as possible, please have the following information ready before reaching out: - The project name affected by the empty DKN. - Confirmation that your knowledge graph completion is at 100%. - The date and approximate time you completed your knowledge graph setup. - Any error messages displayed on the DKN page, if present. - A screenshot of the empty DKN section alongside your completed knowledge graph, if possible. Our team will review your account configuration and trigger or resolve the DKN generation on the backend. If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

🤖 Atlas Agent Stuck on Working State

🔍 What Is This Issue? Atlas Agent may occasionally appear stuck on a 'working on it' message for several minutes without producing a response. This can happen even after reloading the page and is usually related to a temporary service disruption or a background processing delay. ⚠️ Common Causes - Temporary outage: Search Atlas services may be experiencing a brief interruption affecting Atlas Agent's ability to process requests. - High server load: During peak usage periods, response times may be significantly longer than usual. - Stale session: Your browser session may be holding onto an incomplete request that is no longer active. - Network interruption: A brief drop in your internet connection during a request can leave the agent in a pending state. 🛠️ Steps to Resolve the Issue 1. Wait up to 3–5 minutes. Some complex Atlas Agent tasks take longer to complete. If the agent is still unresponsive after this window, proceed to the next steps. 2. Reload the page. Press F5 (Windows) or Cmd + R (Mac) to refresh. Check whether the agent resumes or shows an updated status. 3. Clear your browser cache. A corrupted cache can cause display issues. Clear your cache and cookies, then reload the platform. 4. Try a different browser or incognito window. This rules out browser-specific issues. Supported browsers include the latest versions of Chrome, Firefox, Edge, and Safari. 5. Check your internet connection. Make sure you have a stable connection and are not behind a restrictive firewall or VPN that could interfere with platform requests. 6. Start a new Atlas Agent session. Close the current conversation and open a fresh one. Retype your request to see if the agent responds normally. 📊 How to Check for an Active Outage If none of the steps above resolve the issue, there may be an ongoing service disruption on the Search Atlas side. During an outage, Atlas Agent will be unavailable until the issue is resolved by our engineering team. Outages are typically brief and resolved within a short window. You do not need to take any action on your account — simply try again after a few minutes. 💡 Tips to Prevent This in the Future - Avoid submitting multiple Atlas Agent requests in rapid succession, as this can create queuing delays. - Keep your browser updated to the latest version for the best platform performance. - If you are working on a time-sensitive task, try submitting requests during off-peak hours for faster responses. 🙋 Still Need Help? If Atlas Agent continues to be stuck after following all steps above, our support team can investigate further. If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

📊 Fix Report Recipients and Atlas Agent Issues

📬 Why reports may go to the wrong recipient Reports are sent to the recipient saved in the report or scheduled delivery settings. This address may differ from your Search Atlas account email, especially if the report was created by another user, copied from an older schedule, or configured before your account details changed. 🔍 Check the report recipient 1. Open the report or scheduled report settings. 2. Review the recipient or delivery address. 3. Replace any unfamiliar address with the email associated with the intended Search Atlas user. 4. Save the updated settings and confirm that the next delivery uses the corrected recipient. If you cannot edit the recipient, ask the report owner or an account administrator to update it. Do not rely on the sender address or the address shown in a previous report; check the current delivery settings. ⚙️ Verify your account email Confirm that you are signed in to the correct Search Atlas account and workspace. If your account email was recently changed, create or update the report schedule after the change so the latest recipient information is used. 🤖 Troubleshoot Atlas Agent failures Atlas Agent may sometimes show an incomplete response, report that a connected service is unavailable, or fail to deliver a completed result. First, refresh the page and retry the request with a shorter, more specific prompt. - Confirm that the correct project and report are selected. - Check that required Google Search Console or Google Analytics connections are active. - For large reports, request one section or metric at a time. - Wait for the task to finish before starting another identical request. - If a download is reported as complete, verify that the file actually downloaded before retrying. - For scheduled deliveries, confirm both the destination and the recipient settings. ✅ What to include when retrying State the project, report name, date range, requested action, and expected delivery destination. This helps Atlas Agent use the correct data and reduces incomplete or misdirected results. 💬 Get help from our team If the recipient cannot be changed, Atlas Agent repeatedly fails, or a report still goes to an unknown address, capture the report name and the approximate time of the attempt. If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

🛠️ Reset and Clean Up DKN Nodes

🔍 What Is the DKN and Why Might It Need a Reset? Your Domain Knowledge Network (DKN) is the foundation Search Atlas uses to understand your website's topical structure. It maps your content into interconnected nodes — one per page or topic — to power recommendations across Content Genius and OTTO SEO. Over time, your DKN may develop issues such as: - Duplicate or overlapping nodes — the same page is recognised as multiple distinct nodes. - Stale or poisoned keywords — node keywords carry over incorrect or mismatched data from earlier builds. - Corrupted titles — node titles contain disambiguation code that was not cleaned up automatically. - Mismatched data after a project switch — data from a previous project appears in a new project's DKN. If your DKN contains any of these problems, resetting it and rebuilding from scratch is usually the fastest way to restore accurate topical mapping. ⚠️ Before You Reset: What to Know Please read the following carefully before proceeding: - A full DKN reset is irreversible. All existing nodes, keyword assignments, and topical connections for the project will be permanently deleted. - After a reset, the DKN must be fully rebuilt. Rebuilding time depends on the size of your domain and the number of pages crawled. - Any Content Strategy work tied to existing nodes — such as briefs or cluster assignments — may be affected. Review your active content work before resetting. - Do not use the Sync with Brand Vault button on the Content Strategy page as a workaround for DKN issues. A known issue means this action can wipe DKN nodes without triggering a proper rebuild, leaving your network in an incomplete state. 🛠️ How to Request a Full DKN Reset At this time, a full DKN reset must be performed by the Search Atlas support team. There is no self-serve reset button in the platform. Follow these steps to get it done quickly: 1. Confirm which project needs the reset. Note the exact project name as it appears in Search Atlas. 2. Decide whether you need a full reset (wipe everything and rebuild from scratch) or bulk deletion of specific duplicate nodes (partial cleanup). Both options are supported. 3. Open the chat widget in the bottom-right corner of the platform and type human teammate to connect with the support team. 4. Provide the following information: - Project name and URL - Whether you want a full reset or targeted node deletion - A brief description of the issue you are seeing (e.g., duplicate nodes, corrupted titles, stale data) 5. The team will confirm the action with you before making any changes. 🔄 How to Rebuild Your DKN After a Reset Once the reset is complete, follow these steps to rebuild a clean DKN: 1. Go to your Project Settings and verify that the correct domain and crawl settings are configured. 2. Trigger a new site crawl from the dashboard. This allows Search Atlas to rediscover your pages with fresh data. 3. Navigate to the Content Strategy section and allow the DKN to regenerate automatically as the crawl completes. 4. Once the DKN is rebuilt, review the new nodes to confirm they map correctly to your pages — look for any remaining duplicates or mismatched titles before resuming content work. ❓ Why Are Duplicate Nodes Being Created? Duplicate nodes can appear when the system has difficulty determining whether a page already exists in the network. This can happen in several scenarios: - URL variations — the same page is accessible via multiple URLs (e.g., with and without a trailing slash, or HTTP vs HTTPS), causing the crawler to treat them as separate pages. - Crawl interruptions — a previous crawl did not complete cleanly, leaving partial node data that conflicts with a new crawl. - Disambiguation code leaks — a known bug caused internal code used to differentiate similar topics to appear in node titles and keywords. This issue has been patched, and a backfill was run to clean affected data. If you still see symbols or codes in your node titles, contact support for a targeted cleanup. If you notice duplicates appearing shortly after a rebuild, check your canonical tags and redirect settings in your CMS before running another crawl. Fixing URL consistency at the source prevents the same issue from recurring. ✅ Quick Checklist Before and After a Reset - Note all active briefs and cluster assignments linked to existing nodes. - Confirm the correct project is selected before requesting a reset. - Do not use Sync with Brand Vault as a cleanup workaround. - After the reset, trigger a fresh crawl before reviewing new nodes. - Check canonical tags and redirects in your CMS to prevent duplicate nodes from reappearing. 💬 Still Need Help? If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

Fix Stuck Microagent Tasks and Common Bugs

Overview Some customers have reported a combination of issues affecting Microagent tasks, Instant Indexing setup, and Content Genius features. This article explains each known issue and sets accurate expectations for what to have ready when escalating to the support team. Stuck Microagent Tasks in PENDING Status Microagent tasks can become stuck in a PENDING state and never complete. If you see tasks marked PENDING that are not progressing, note the following details before escalating: - The name of the affected project or Content Genius document - The approximate time the task was created and how long it has been stuck - Any error messages or status indicators you see on screen Our team will need these details to investigate and manually reprocess affected tasks on the backend. Blank Service Account Email for Instant Indexing When activating Instant Indexing, some users encounter a blank service account email field. This leaves the Google Search Console (GSC) verification step incomplete and can cause the indexing status to appear contradictory. If your service account email field is still blank, this requires manual provisioning by our team. Before escalating, please have the following ready: - Your project or site name in OTTO SEO - A screenshot showing the blank service account email field - Confirmation of which GSC property you are trying to connect Known Content Genius Bugs Some reproducible bugs affecting Content Genius features have been identified. If you are experiencing issues with Content Genius — such as problems with keyword management, SEO scoring, or content generation — please escalate with the following information: - The name of the affected project and document - A description of the exact behavior you are seeing - The exact error message, if any, along with the date and time it occurred If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

Fix Missing Projects in Agent Hub Portfolio List

Overview If a project is visible in other parts of the platform but missing from your Agent Hub portfolio list, you are not alone. This is a known sync issue that can affect projects in your account. Your project data is not lost — this is a display issue between the two interfaces. Why This Happens The Agent Hub portfolio list and other platform views pull project data differently. In some cases, a project may not immediately surface in the Agent Hub portfolio list, even when it is fully configured and working elsewhere in the platform. This is a sync gap between the two interfaces, not a loss of data or project configuration. What to Do If Your Project Is Missing Because this is a backend sync issue, our support team will need to look into your specific account to resolve it. To help us do that as quickly as possible, please have the following ready when you reach out: - The exact name of the project that is missing from Agent Hub - Confirmation of whether the project is visible in other parts of the platform - The approximate date and time you first noticed the project was missing - Your account email or workspace name What Will Not Be Affected This issue is a display sync problem. Your underlying project data and configurations remain intact regardless of whether the project appears in the Agent Hub portfolio list. If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

🔍 OTTO Project Missing from Atlas Brain Index

Overview If your OTTO project appears to exist in the backend but is missing from the Atlas Brain search bar or project list, and attempting to re-add it returns a Request failed: 400 error, this is a known indexing issue. Because the project already exists in the system, the platform blocks duplicate additions — but the project also fails to surface correctly in Atlas Brain. This requires a manual fix from the Search Atlas team. What Causes This Issue This issue occurs when an OTTO project becomes out of sync between the backend and the Atlas Brain search index. The project record exists in the database, so re-adding it triggers a 400 error (duplicate entry), but it is not registered in the index Atlas Brain uses to display and search projects. This prevents Atlas Brain from connecting to the project or delegating tasks to it correctly. How to Resolve It Because this issue involves a backend index mismatch, it cannot be fixed through the platform interface alone. To get the issue resolved as quickly as possible, follow these steps: 1. Confirm the project exists: Verify that your OTTO project (for example, brittanyelise.com) is visible in your OTTO section and is listed as active. Make a note of the exact domain or project name. 2. Do not attempt to re-add the project: Trying to add the same project again will return a Request failed: 400 error because the backend already has a record of it. Additional re-add attempts will not resolve the issue and may complicate the fix. 3. Gather the following information before contacting support — having these details ready will allow the team to locate and re-register your project in the Atlas Brain index as quickly as possible: - The exact domain or project name (for example, brittanyelise.com) - A screenshot or description of the 400 error you received when attempting to re-add the project - Your account email or workspace identifier 4. Submit a support request by opening the chat widget in your Search Atlas dashboard (click the chat icon in the bottom-right corner) or emailing the support team directly. Paste the information you gathered in the previous step into your message so the team can prioritize and action your request without needing a follow-up. 5. What happens next: Once the support team receives your details, they will manually re-register the project in the Atlas Brain index. This resolves the mismatch and makes the project appear correctly in Atlas Brain search and the project list. Most cases are resolved within one business day. If you do not hear back within 24 hours, reply to your original support thread to follow up. If you have any questions or need further assistance, please contact our support team and we'll be happy to help.

Last updated on Oct 26, 2026

🔄 Fix Account Sync Issues and Golden Loops

Overview Some Search Atlas users experience account-wide sync issues where changes made by one admin are not visible to others, or where Golden Loops fail to save correctly. This article explains what to do if you encounter these problems. Step 1 — Confirm Your Workspace Access Before troubleshooting sync, verify that all team members are in the same workspace and have the correct roles assigned. Check your workspace settings to confirm that any admins who should share access are listed with the appropriate permissions. If a team member is missing, invite them and ask them to accept the invitation before testing sync again. Step 2 — Refresh and Re-authenticate Sync issues are sometimes caused by a stale browser session. Try the following before escalating: 1. Save any unsaved work. 2. Log out of your Search Atlas account completely. 3. Clear your browser cache and cookies, or try an incognito/private window. 4. Log back in and navigate to the affected area. 5. Ask the other admin to do the same on their device, then check whether changes are now visible to each other. Step 3 — Check Golden Loop Settings If a specific Golden Loop is not saving or is behaving unexpectedly, navigate to the Golden Loops section of your account. Open the Golden Loop that is not saving correctly, re-select any required settings such as the associated Project, and save again. If the issue persists after saving a second time, note the exact details of what is not persisting so you can share them when escalating. When to Escalate If the steps above do not resolve your sync issue or your Golden Loop continues to fail to save, please have the following information ready before contacting support: - Your workspace or project name - The exact behavior you are seeing (e.g., which settings are not saving, which admin accounts are affected) - Any error messages displayed on screen - The approximate time the issue first occurred If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

🧠 Fix Atlas Brain Fetch Failures

🧠 What “Failed to fetch” means The “Failed to fetch” message usually indicates a temporary connection problem between Atlas Brain and its processing services. It does not necessarily mean your project, prompt, or account is damaged. 🔄 Try these quick fixes 1. Wait a few minutes, then submit your request again. 2. Refresh the Search Atlas platform and reopen Atlas Brain. 3. Check that your internet connection is stable. 4. Try a shorter prompt if the original request included a large amount of text, files, or images. 5. If you were using a connected tool or data source, reopen it and try again. 📚 If your history is missing A temporary loading or service error can prevent recent conversations from appearing. Refresh the page, confirm that you are in the correct workspace, and check again after a few minutes. Avoid deleting or recreating the project while the service is recovering. If the conversation still does not appear, start a new chat and briefly describe the work you were completing. Keep a copy of important prompts and responses outside Atlas Brain while troubleshooting. 🖼️ Requests involving images or tools Atlas Brain may temporarily fail when analyzing images, connecting to backend tools, or delegating work to another agent. Retry the request without the attachment or tool step first. If that works, add the image or tool request again in a separate message. ✅ When to contact support If the error continues after refreshing and retrying, note the approximate time, the request you submitted, and any connected tool or file involved. Do not include passwords or sensitive information. If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

🛠️ Fix AI Agent Edits on Pillar Blog Pages

🔍 Overview Some customers have reported that the Website Studio AI agent declines to make edits — such as adding internal links — to pillar blog pages, even when the same prompt worked successfully before. This article explains what is known about this issue and what to have ready when reaching out for support. ⚠️ Why This Happens This is a known bug affecting the Website Studio AI agent when working on pillar blog pages. The root cause has not been fully confirmed, but reported symptoms include the agent refusing edit actions or failing to respond as expected on pillar page types. Because the underlying cause is being investigated on the backend, there are no verified self-serve steps that reliably resolve the issue. 📋 What to Have Ready When You Escalate To help our team investigate and resolve this as quickly as possible, please have the following information ready before reaching out: - Project name: The exact name of the Website Studio project where the issue occurred. - Pillar page URL: The specific page URL where the AI agent refused to make edits. - Prompt used: The exact prompt or instruction you entered into the AI agent when the failure occurred. - What you expected vs. what happened: A brief description of the intended edit and the agent's actual response (e.g., error message, refusal text, or no response). - Timestamp: The approximate date and time when the issue occurred, including your timezone. If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be connected with a member of our team.

Last updated on Oct 26, 2026

💡 Fix Identical AI Review Responses

🔍 Why responses may repeat AI-generated review responses can sound identical when reviews share similar wording, ratings, or topics. Repeated instructions, short reviews, and limited business details can also lead to similar suggestions. 🛠️ Create a more specific response 1. Go to Local → GBP Projects in the left sidebar. 2. Open the relevant project with View project. 3. Open Overview → Reviews. 4. Review the generated response and edit it before publishing. 5. Add details from the customer’s review, such as the service used, staff member mentioned, or specific concern. ✍️ Improve response variety - Start with different openings instead of using the same greeting each time. - Refer directly to the customer’s words and experience. - Vary the structure: thank the customer, address the feedback, and close with an invitation to return. - Use a different tone when appropriate, such as warm, apologetic, appreciative, or solution-focused. - Avoid adding personal details that are not included in the review. ✅ Review before publishing Always check each suggestion for accuracy, tone, and uniqueness. AI-generated content is a starting point, so make final edits to ensure the response sounds natural and reflects the specific customer interaction. 💬 Get help If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be instantly connected with a member of our team.

Last updated on Oct 26, 2026

🛠️ Understanding Project Issue Status and Deploying Fixes

🔍 Why This Question Comes Up Many customers ask, "Is the issue on my project resolved yet?" or "Are there any updates?" Most of the time, the answer depends on one thing: whether the fixes the platform generated for you have been deployed. This article explains how issue status works, what to do first, and how to check that an issue is actually closed. 📌 Generated Fixes vs. Deployed Fixes These two steps are easy to confuse, but they are not the same thing: - Generated: The platform has identified a problem on your site and prepared a suggested fix. Nothing on your website has changed yet. - Deployed: The fix has been applied so it is live on your website. An issue will generally stay open until its fix is deployed and the platform has had a chance to check your site again. A generated fix on its own does not resolve the issue. 🚀 Step 1: Deploy Your Generated Fixes First Before asking for an update on an issue, make sure the fixes for it have been deployed. Follow these steps: 1. Open your project in the platform. 2. Go to the section that lists your project issues or recommendations. 3. Find the issue you are concerned about. 4. Review the generated fix to make sure it looks right for your site. 5. Deploy the fix using the deploy option shown for that item. 6. Wait for the platform to confirm that the deployment finished. If you have several generated fixes waiting, deploy them first. Then check the status of your issues again. This often answers the question on its own. 📊 Step 2: Check the Issue Status Once your fixes are deployed, look at the status of the issue in your project. Here is what the common states usually mean: - Open or pending: The issue is still detected, or the fix has not been deployed yet. - Fix generated: A suggested fix is ready, but it still needs your review and deployment. - Deployed: The fix has been applied. The platform may still need time to re-check your site. - Resolved: The platform no longer detects the issue on your site. Exact labels can vary slightly depending on the tool you are using, but the idea is the same: deployed first, then verified, then resolved. ⏳ Step 3: Allow Time for Re-Checking After you deploy a fix, the issue may not change to resolved right away. The platform needs to look at your site again to confirm the problem is gone. Keep these points in mind: - Give the platform some time to re-scan your project after deployment. - If you have a rescan or re-audit option, run it to refresh the results. - Refresh your browser page to make sure you are seeing the latest status. 🧩 Why an Issue Might Still Show as Open If you deployed a fix and the issue still appears, one of these may be the reason: - The fix was not fully deployed. Check that the deployment completed without errors. - The project has not been re-checked yet. Run a new scan or wait for the next one. - Your site is serving an older version. Caching on your website or hosting can delay changes from showing. Clearing your site cache can help. - The issue exists on more than one page. Fixing one page does not close an issue that affects several pages. - A different problem was found. A new issue may look similar to the old one but have a different cause. ✅ How to Confirm an Issue Is Resolved 1. Confirm the fix shows as deployed in your project. 2. Run a fresh scan or wait for the next scheduled one. 3. Reopen the issue list and check that the item is marked resolved or no longer appears. 4. Visit the affected page on your live site to make sure the change is visible. 💡 Tips to Avoid Delays - Deploy generated fixes as soon as you have reviewed them. - Work through issues in batches, then rescan once. - Keep a note of which issues you are tracking so you can check them quickly. - When you ask for help, mention the project name and the specific issue. This helps our team find it faster. 💬 Still Need Help? If you deployed your fixes, allowed time for a re-check, and the issue still looks unresolved, we are happy to take a closer look. If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be instantly connected with a member of our team. To get the quickest answer, include the name of your project, the issue you are asking about, and whether you have already deployed the generated fix.

Last updated on Oct 26, 2026

Fix Outdated or Incorrect Fin Responses

🔍 Check the response first Fin responses are generated from available Search Atlas information and may occasionally be outdated, incomplete, or incorrect. Before taking action, compare the response with the current platform interface and documentation. 1. Confirm that you are viewing the correct Search Atlas project and feature. 2. Navigate directly to the relevant section of the platform and check the current settings, options, or page layout for the latest information. 3. Review the response for missing context, incorrect steps, or references to features that may have changed. 🧭 Provide more context If Fin misunderstood your question, ask again with specific details. Include the feature name, your goal, the steps you followed, and what you expected to happen. You can also include the exact wording of the response and the page where the issue occurred to help clarify the context. 🛠️ Cross-check with the current platform If a Fin response describes a navigation path, button, or setting you cannot locate, do the following: 1. Open the relevant tool directly within the Search Atlas platform and compare what you see with what Fin described. 2. If the feature or option Fin mentioned does not appear in the current interface, treat the response as potentially outdated and do not act on it for important decisions. 3. Check any in-platform tooltips or help icons near the relevant section for up-to-date guidance. 💬 Report an incorrect or outdated response If Fin continues to provide incorrect or outdated information, report it so the team can review and improve future answers. When escalating, have the following ready: - The exact Fin response that appears incorrect or outdated. - The feature or section of the platform where the issue occurred. - A description of what the correct behavior or information should be, if known. - Any screenshots or page URLs that illustrate the discrepancy. Do not rely on an outdated Fin response for important account, SEO, or implementation decisions until the information has been verified directly in the platform. If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be instantly connected with a member of our team.

Last updated on Oct 26, 2026

🤖 Fix Unfinished Playbook Names

🔎 What happened An unfinished working note may have been saved or published as the name of a playbook in the Playbook Library. This can make an internal draft appear as a customer-facing or reusable playbook title. 🛠️ What to check 1. Open the Playbook Library. 2. Find the playbook with the unfinished note as its name. 3. Open the playbook and review its title, instructions, and status. 4. Confirm whether the content is a draft, an active playbook, or an accidental publication. ✏️ Correct the playbook 1. Rename the playbook with a clear, descriptive title. 2. Remove unfinished notes, placeholders, and internal comments from the playbook content. 3. Save the updated playbook. 4. If the playbook was published accidentally, unpublish or archive it when that option is available in your workspace. 5. Refresh the Playbook Library and confirm that the corrected name appears. ✅ Prevent the issue - Use a clear title before saving a new playbook. - Keep working notes separate from final playbook instructions. - Review the title and content before publishing. - Use draft status while the playbook is still being edited. 🚨 If the issue continues If an unfinished working note is published again, or you cannot rename, unpublish, or archive the affected playbook, record the playbook name and the steps that caused the issue. Avoid deleting the playbook until the issue has been reviewed. If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be instantly connected with a member of our team.

Last updated on Oct 26, 2026

🛠️ How to Recover a Deleted Conversation

Deleting a conversation by accident is easy to do, especially when you are cleaning up a long list. The good news is that many customers have been able to find their conversation again on their own. This article walks you through what to check first, what to do if you cannot find it, and how to avoid the same mistake in the future. ❓ What Happens When You Delete a Conversation When you delete a conversation, it is usually removed from your conversation list right away. Depending on how it was removed, it may or may not be recoverable. That is why it helps to act quickly and avoid making changes that could overwrite or hide it further. - Deleting a conversation removes it from your main list. - Starting new conversations will not remove your other ones, but it can make the missing one harder to spot. - If you are unsure whether it was deleted or just hidden, the checks below will help you tell the difference. 🔍 Step 1: Check Whether It Is Really Gone Before assuming the conversation is lost, rule out the simple causes. Many conversations that look deleted are actually just out of view. 1. Refresh the page or reopen the platform in a new browser tab. 2. Scroll through the full conversation list, not just the most recent items. 3. Look for a search bar or filter and clear any filters that may be hiding older items. 4. Check whether you are in the correct project, workspace, or account. Conversations are tied to where they were created. 5. Look for any archive, history, or trash section in the area where you manage conversations. ♻️ Step 2: Try to Restore It Yourself If you find the conversation in an archive, history, or trash area, you can usually bring it back without any help. 1. Open the section where the conversation appears. 2. Select the conversation you deleted by mistake. 3. Choose the option to restore, unarchive, or move it back to your main list, if one is shown. 4. Return to your main conversation list and confirm it appears with its full history. Several customers have recovered their conversation this way, so it is always worth checking before reaching out. 🔄 Step 3: Try a Fresh Session Sometimes the platform is showing out-of-date information. If you still cannot see the conversation, try the following. - Log out of the platform and log back in. - Clear your browser cache or try a different browser. - Open the platform in a private or incognito window to rule out browser extensions. - Confirm you are signed in with the same account you used when the conversation was created. 📝 Step 4: Gather What You Remember If the conversation still cannot be found, gather a few details. This makes it faster for our team to help you. - The approximate date and time you deleted it. - The name of the person or the topic of the conversation. - The project or workspace you were working in. - Any key information from the conversation that you need back, such as links, recommendations, or notes. If you copied any part of the conversation elsewhere, such as a document or email draft, check there as well. You may already have what you need saved. 🚀 If You Only Need the Information Back Sometimes the goal is not the conversation itself but the answers inside it. If you cannot restore it, you can often rebuild what you need quickly. - Start a new conversation and summarize what you remember about the topic. - Ask the same questions again to get fresh results. - Check related reports, dashboards, or project pages, which may hold the same data. 💡 How to Avoid Deleting a Conversation by Mistake - Slow down on cleanup. Double-check the name of the conversation before you confirm deletion. - Read the confirmation message. If a prompt appears, make sure it refers to the conversation you intend to remove. - Archive instead of delete. If an archive option is available, use it for conversations you might need later. - Save important details. Copy key information into your own notes or documents. - Use clear names. Descriptive conversation names make it easier to pick the right one. 🙋 Frequently Asked Questions Can a deleted conversation always be recovered? Not always. It depends on how the conversation was removed. Check for an archive or history area first, then contact our team if you cannot find it. Will deleting a conversation affect my other work? No. Removing one conversation does not affect your other conversations, projects, or reports. Does a new conversation replace the one I deleted? No. A new conversation is separate and will not bring back the deleted one. How soon should I act? As soon as possible. The sooner you check for the conversation or ask for help, the better the chance of finding it. 💬 Need More Help? If you have tried the steps above and still cannot find your conversation, our team is happy to take a look. If you need further assistance, open the chat widget in the bottom-right corner of the platform and type human teammate to be instantly connected with a member of our team.

Last updated on Oct 26, 2026