What is Knowledge?
Knowledge is our RAG solution in Relevance AI, which allows you to provide additional context and relevant information to your agents and tools.
Create a Knowledge table
There are a couple of ways to create a knowledge table.
- Blank: Start with an empty table and manually input your data.
- Click on “Blank” and name your table. Click on “Add column”, then “Blank column”, then enter a name, select a column type and create your column. Click “Add row” to enter data.
- Upload from CSV: Import structured data from a CSV file.
- Select "Upload from CSV". Choose your CSV file and hit “Upload data to table”.
- Import from website: Automatically extract and structure data from a specified website.
- Enter the name of your table, and URL of the website you want to import from. Select the type of content to extract (text or HTML).
- Import from 3rd Party: Connect to third-party services to import relevant data.
- Choose from available third-party integrations.
- For example, if you select “Slack”, then enter your API key, select the channel you want to important from, and the number of messages.
How to access Knowledge
To give your tool access to knowledge, select the “Knowledge Search” tool step. Then, select the knowledge table you want to search, type in your query and select between “keyword” (search for a specific word) or “vector” (search for relevant information).
To give your agents access to knowledge, make a tool, add the Knowledge Search step, and give your agent your tool.
Enrich with tool
Enriching with tools allows you to enhance your knowledge base using AI tools.
Select the knowledge table you want to enrich, click on ‘Enrich with Tool” and choose the tool your want to use from the library.
Configure how the tool should process and add information to your table by mapping outputs to columns, then run the enrichment.