How do I configure my workspace's AI?
Open Settings, then the AI Agent tab. From there you see which model answers at the workspace level (it's presented simply as diagnos), and you turn on the connection that lets AI tools you already use — ChatGPT, Claude, anything that speaks MCP — send messages and files into your workspace's agents.
- Updated on
- 2 min read
On this page
Workspace-level AI settings live under Settings, in the AI Agent tab. From there you see which model answers at the workspace level, and you turn on the connection that lets AI tools you already use — ChatGPT, Claude, or anything else that speaks MCP — send messages and files into your workspace's agents.
Connecting an outside AI tool is a one-time setup, like connecting any other third-party app. Once it's connected, anyone using that tool can address one of your workspace's agents directly — for example, "send this to Red 1" — without opening diagnos at all.
1. Open AI settings#
Go to your workspace's Settings, then the AI Agent tab.
2. Model#
Inside an exam, every agent answers with the same fixed model, so clinical answers stay consistent no matter who's asking — there's nothing to choose there. At the workspace level, the assistant is presented to you simply as diagnos; there's one option behind that name today, with room to add more later without changing how it's presented to you.
3. Connect outside AI tools#
Turning on the connection lets AI tools you already use reach your workspace's agents over the same open protocol most AI apps already speak (MCP). It can only send things in and read diagnos's public help center — see What can the AI assistant help with? for exactly what it can and can't reach.
Before you turn it on#
Same rule as every AI conversation: never share sensitive data such as names, emails, or phone numbers there — it's an AI, and it can make mistakes. AI conversations aren't part of diagnos's end-to-end-encrypted vault.