Guides

Document Analyst - Overview

Document Analyst

Document Analyst is a CustomGPT.ai feature that lets users analyze documents during conversations. Users can upload a file directly in chat, or ask the agent to read a specific document already stored in your knowledge base. In both cases, the agent analyzes the document against your knowledge base to deliver contextual responses.

Two modes

Upload & Analyze

Users attach a file in the chat input and ask questions about it. The agent reads the file and combines its content with your knowledge base to generate a response. Files are available for the duration of the session only and are not added to your knowledge base.

Supported file types: pdf, docx, doc, odt, txt, jpg, jpeg, png, webp

For image files, the agent uses AI vision to interpret the content - reading diagrams, screenshots, or scanned pages the same way it reads text documents.

Analyze from knowledge base

The agent reads a specific document already in your knowledge base from start to end, rather than searching across fragments. Users reference the document by name in their prompt - for example, "Summarize the August Report" or "Tell me more about that research document you mentioned."

The agent detects the document reference, locates the file, and performs a full read. If the document cannot be found, the agent asks the user to clarify.

For knowledge base documents that are images (jpg, jpeg, png, webp), the agent automatically processes them with AI vision at query time. No configuration is needed.

Availability

Document Analyst is available on Standard, Premium, and Enterprise plans. It works across web, mobile, embedded widgets, Slack, and API deployments.

Related articles


Updated 8 days ago