What can go into a HighLevel Knowledge Base?
HighLevel’s current Conversation AI training documentation lists website pages, FAQs, documents, tables, rich text, Google Drive files and other supported sources. The objective is to give agents business-specific information rather than relying on generic model knowledge.
Build for retrieval, not for storage
A Knowledge Base is not an archive. Include information the agent actually needs to answer, qualify or act. Exclude outdated policies, duplicated service descriptions, navigation text and irrelevant legal boilerplate unless they answer a real customer question.
Source-type strategy
| Source | Best use |
|---|---|
| Website pages | Current public services, policies and pricing explanations |
| FAQs | Short, approved answers to frequent questions |
| Tables | Structured service, product or pricing information |
| Files / Drive | Approved internal reference documents that need retrieval |
| Rich text | Business rules or instructions maintained directly in HighLevel |
Testing has two layers
HighLevel distinguishes retrieval testing from full bot testing. First confirm that the correct source information can be found. Then test whether the agent uses that information correctly inside a realistic conversation with its prompts, goals and actions.
Maintenance is part of the system
Knowledge degrades when prices, hours, staff, services or policies change. HighLevel supports refreshing web sources, and agencies should assign ownership for updating non-web sources too. A reliable AI agent is partly a content-maintenance process.
Try HighLevel AI
Review HighLevel’s current AI offer, product details and trial terms on the merchant site.
Try HighLevel AI ↗Affiliate disclosure: we may earn a commission if you sign up through this link, at no extra cost to you.
Sources & verification
Mutable product details were checked against primary vendor documentation. Fact-check date: September 13, 2026.