Ways to create Knowledge
Empty table
Start from scratch and structure the data yourself — useful for FAQ-style or tabular content.
Upload files
Import CSV, PDF, Excel, JSON, or audio files. Relevance AI parses them into a searchable table.
Website
Extract content from one or more URLs — handy for company sites and documentation.
Integrated source
Sync from Google Drive, SharePoint, Notion, or Confluence so your Knowledge stays current.
- Click Knowledge in the left sidebar.
- Click Create Knowledge.
- Pick how you want to start, then follow its prompts:
- Empty table — name the table and add columns (for example Question, Answer, Category), then enter rows directly.
- Upload files — choose your file(s), confirm the parsing options, and name the table. Supported formats are CSV, PDF, Excel, JSON, and audio.
- Website — enter the URL(s) to crawl and name the table.
Integrated Knowledge sources
Connect a third-party source so your Agents always read the latest version of your documents. Each source has its own setup guide:Google Drive
Read selected Google Drive files as Knowledge.
SharePoint
Read selected SharePoint files and folders as Knowledge.
Notion
Read selected Notion pages and databases as Knowledge.
Confluence
Read selected Confluence pages and spaces as Knowledge.
Connect Knowledge to an Agent
- Open your Agent and go to the Build screen.
- Click Knowledge, then Add existing Knowledge.
- Select the Knowledge set you want to add.
- Choose how the Agent should use it:
- Add all to prompt — includes the entire Knowledge set in the prompt every time the Agent runs. Best for small or simple datasets.
- Allow agent to search — the Agent searches the Knowledge set with RAG and pulls only what’s relevant. Best for large or complex datasets.
What’s next?
- Enrich with a Tool — run a Tool across your table to add summaries, tags, or derived fields.
- Knowledge Tool steps — read from and write to Knowledge inside a Tool.
- Delete knowledge — remove a Knowledge table when you no longer need it.

