Skip to main contentKnowledge bases
The brains behind your AI Agents.
In order for your AI Agents to know how to help your customers, they need to have a bank of structured information to work with. This information is called a knowledge base.
Information in a knowledge base is stored in documents, which are nested within folders.
Example knowledge base folder structure
Take a look at the drawing below. It shows you how a section of a knowledge base for an ecommerce store called ‘Dream Cosmetics’ that sells beauty products might be structured.
The folders are in blue. The documents are in yellow.
Most knowledge bases will be more complex than this example, but the basic principles of organization will remain the same, even when scaled up:
- Organize your topics from most broad to most specific.
- Group related topics together.
- Make sure your topics don’t overlap.
Example knowledge base document structure
The information stored in your documents also needs to be carefully structured to see the best results from your AI Agent.
What content to include
The content that you add to your knowledge base is completely up to you. The more information you add, the more helpful your AI Agent will be. Some examples include:
- FAQs
- Ordering and shipping information
- Return and exchange policies
- Account management
- Payment and billing information
- Store locations
- Company background and information
Product information
We recommend that you don’t use Knowledge Bases to provide your AI Agents with product information. Instead, connect a Shopify store to your workspace and use the ‘Recommend products’ task for best results.
Tips and tricks
Here are some ways you can set your knowledge base documents up for success:
- Use headings, bullet points, and numbered lists to signpost your content.
- Provide information in full sentences.
- Use simple language.
- Structure your documents with clear headings and sections.
- Include examples where helpful.
- Keep information up-to-date and relevant.
