Skip to content
Your AI agent

Knowledge

Documents your assistant reads to answer customers' questions.

1 Show a source line in replies. 2 Upload a document. Screenshot taken 10 November 2026.

Overview

Knowledge is for the longer material your assistant should know but that doesn't fit in FAQs or Services: policies, product guides, price sheets, internal notes. Upload the documents and the assistant searches them when a customer asks something they cover. You can keep up to 100 files, 10 MB each. You find it under Settings, in Knowledge.

Features

Uploaded Files

Every document the assistant can use, with buttons to edit its description and citation reference, download it, or delete it. A counter shows how many of your 100 files you've used.

Upload files

Drag a file onto the upload area or click it to choose one. PDF, DOC, DOCX, PPTX, TXT, MD, HTML, JSON and code files are supported, up to 10 MB each. Before you click Upload, the AI writes a short Description of the file for you; change it if it isn't quite right.

Cite source documents in responses

When on, a reply that uses one of your documents ends with a "Source:" line showing that document's citation reference and a short quote. Only documents with a Citation reference are ever cited. This switch saves straight away.

Citation reference (optional)

The name shown in the "Source:" line when the assistant cites this document, for example "Patient Guide 2026". Leave it blank to keep that document out of citations.

Common tasks

Add a knowledge document

About 1 min
  1. In the sidebar, open Agent Flow (called Agent Builder on some accounts), click Settings, then choose Knowledge. If your account also has the Coordinator, pick Chat agent first.
  2. Drag a file onto the upload area, or click it to pick a file.
  3. Check the Description, add a Citation reference if you want, and click Upload. The file then appears in Uploaded Files.

Tips

Split long documents into smaller files by topic. Smaller, focused files are easier for the assistant to search.
A clear description helps the assistant understand what is in each file.

Frequently asked questions

Was this page helpful?