Plannor for ChatGPT

AI Interaction Disclosure

Plannor aims to give teams a calm, supportive AI experience without relaxing the safety rules that protect tenant data, role boundaries, and operational accountability.

How the ChatGPT Connection Works

After a user connects their account, ChatGPT can call approved Plannor tools through the MCP endpoint. Each call is resolved inside Plannor using the authenticated user, tenant, role, and permitted scopes.

Tool access is role-aware and tenant-scoped.
Tool calls are audited.
We do not trust tool inputs for identity or tenant selection.

Model and Credit Usage

ChatGPT MCP calls execute Plannor business logic directly and do not consume Gemini API credits in this architecture.

The built-in Plannor chatbot still uses Gemini for text and live responses.
The MCP connector reuses the same business logic layer and tool catalog, not the Gemini text-generation path.
This separation helps keep ChatGPT connector costs predictable.

What Users Should Expect

AI output can be helpful, but it should still be reviewed by a human for important decisions. Users should verify sensitive records, billing actions, care operations, and administrative changes before relying on them.

Use the connector for approved operational workflows.
Expect some tools to require clearer user intent before acting.
Review changes made through AI in your normal Plannor audit and reporting flows.