Skip to main content

Search Knowledge Base by Keyword

This KB documentation is using Advanced Search, Links Editor and Elegant Layouts add-ons.

Print

Step 2 – Add AI Training Data

The AI Chat and AI Search are designed to assist your users, but to truly understand your business or organization, it needs to be trained.

This training process involves feeding the AI relevant materials such as:

  • KB Articles
  • Posts
  • Pages
  • PDFs
  • FAQs
  • Notes

Private Articles and Private Posts cannot be used for AI Search and AI Chat. Only public posts, public articles, AI Notes, and PDFs are supported.

By doing so, the AI can learn the nuances of your workflows, terminology, and customer needs, enabling it to provide accurate and tailored responses.

Think of it like onboarding a new team member—without the right information, the AI can only offer generic answers. Once trained with your materials, it becomes a knowledgeable partner, capable of addressing specific inquiries and supporting your team or customers more efficiently.

AI Training Data Support

Access Manager does not handle AI Training Data protection.  Therefore, Private articles cannot be submitted for AI training.

Start small
– Begin with 10–20 of your best articles. Test chat/search quality before uploading the full library — it’s faster to iterate and cheaper on tokens.

Quality over quantity
– Only include articles that are complete, accurate, and up to date. Outdated or half-written content trains the AI to give outdated or half-written answers.

Write for answers, not SEO
– Lead with the direct answer, then explain. The AI pulls from the top of articles most heavily — bury the answer and it gets missed.

Use clear titles and headings
– Titles like “How to reset your password” beat “Password Tips.” Descriptive H2/H3 headings help the AI segment content correctly.

One topic per article
– Split long “everything about X” articles into focused pieces. Retrieval works better when each chunk is about a single thing.

Remove noise
– Strip promotional blurbs, “click here” links, repeated boilerplate, and navigation text. The AI treats this as content and it dilutes relevance.

Keep terminology consistent
– Pick one term per concept (“customer” vs “client” vs “user”) and stick with it. Inconsistency makes the AI hedge or return mixed results.

Add synonyms and alternate phrasings
– If users search “refund” but your article says “return policy,” include both terms somewhere in the article.

Re-sync after edits
– Training data is a snapshot. When you update articles, re-sync the collection so the AI sees the new version.

Was this article helpful?
0 out Of 5 Stars
5 Stars 0%
4 Stars 0%
3 Stars 0%
2 Stars 0%
1 Stars 0%
5
How can we improve this article?
Please submit the reason for your vote so that we can improve the article.
Need help?
Table of Contents