🗺️ Overview
The Doorbell Knowledge Network (DKN) content pipeline in Search Atlas lets you generate SEO articles at scale. Before those articles go live in GoHighLevel, you may want to automatically apply a custom refinement prompt — for example, adding an About the Author section, a Contact Us block, or stripping unwanted formatting artifacts. This article explains how to configure that automation so every article is polished and deployment-ready without manual editing.
✅ What You Can Automate
A custom refinement prompt can instruct the AI to perform any combination of the following tasks on each generated article before it is pushed to GoHighLevel:
- Add a branded About the Author section at the end of the article
- Insert a Contact Us call-to-action block
- Remove formatting artifacts such as extra line breaks, stray markdown symbols, or duplicate headings
- Enforce a consistent tone, reading level, or word count
- Append custom disclaimers, schema hints, or internal linking instructions
🛠️ Step-by-Step: Setting Up Your Refinement Prompt
- Open your DKN project. From the Search Atlas dashboard, navigate to Content and select the DKN project you want to configure.
- Go to Project Settings. Click the Settings (gear) icon within the project to open the configuration panel.
- Locate the Content Refinement section. Scroll to the Post-Generation Prompt or Content Refinement field. This is where your custom instructions are entered.
- Write your refinement prompt. Enter the exact instructions you want the AI to follow on every article. Be specific and use clear directives. See the example prompt below for guidance.
- Save your settings. Click Save to apply the prompt to the project. All subsequent articles generated in this project will automatically pass through your refinement instructions before deployment.
- Connect your GoHighLevel destination. If you have not already linked your GoHighLevel account, go to Avatar (top-right) → Settings → Integrations in Search Atlas, authenticate your GoHighLevel workspace, and map it as the publish destination for this DKN project.
- Run a test article. Generate a single test article and preview the output to confirm the refinement prompt is applying your changes correctly before enabling bulk deployment.
✍️ Example Refinement Prompt
Copy and adapt the following prompt to match your brand requirements:
Prompt:"After generating the article, perform the following refinements: (1) Remove any stray asterisks, hashtags, or markdown formatting symbols. (2) Add a section titled 'About the Author' immediately before the closing paragraph, using the bio provided in the project settings. (3) Add a 'Contact Us' section at the very end of the article with the following text: [insert your contact copy here]. (4) Ensure the article contains no duplicate headings and that all H2 and H3 tags follow a logical hierarchy. Do not alter the core SEO content or keyword placement."
Adjust the numbered instructions to include only the steps relevant to your workflow.
⚙️ Tips for Writing Effective Refinement Prompts
- Be explicit. Tell the AI exactly where to place new sections (e.g., "immediately before the last paragraph" rather than "somewhere near the end").
- Use numbered steps. Numbered instructions reduce ambiguity and ensure every task is acted on in order.
- Protect your SEO work. Include a line such as "Do not change keyword usage, meta description, or heading structure" to prevent the refinement pass from undoing your optimisation work.
- Store reusable copy in Project Settings. If your Author Bio or Contact Us text is long, save it as a project-level variable so the prompt stays concise and easy to maintain.
- Test before bulk deployment. Always validate the refined output on one article before scheduling a large batch to GoHighLevel.
❓ Frequently Asked Questions
Does the refinement prompt apply to articles already generated?
No. The post-generation prompt applies only to articles generated after you save the new settings. To refine previously generated articles, you would need to regenerate them or manually edit them.
Will this slow down article generation?
There is a small additional processing step, but for most projects the impact on generation time is negligible.
Can I use different prompts for different DKN projects?
Yes. Each DKN project has its own independent prompt settings, so you can customise the refinement instructions per project or per client.
What if my GoHighLevel articles still show formatting issues after the prompt runs?
Check that your refinement prompt explicitly names the artifact to remove (e.g., "remove all double asterisks"). Vague instructions like "clean up formatting" may not catch every case. Refine your prompt and run another test article.
💬 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.