Your AI
Bring your AI. MIS does the rest.
MIS has an MCP connector: the assistant you already use — Claude, ChatGPT, Gemini — connects to the platform and answers from your data, not by guessing. MIS stays the perimeter: view, act, or nothing — you decide, user by user.
How it works
An open protocol, not a padlock.
MCP (Model Context Protocol) is the open standard AI assistants use to connect to the tools you already work with. MIS already speaks it: connect your assistant and ask what you need in one sentence. No new assistant to learn, no extra AI cost: you already pay for your AI — MIS gives it access to your data.
What you ask
Real questions, answers from your data.
What equipment is free on September 15th?
Where does the October conference production stand?
How much did I invoice on June's productions?
Permissions
The AI sees what you see. Nothing more.
AI access starts disabled — by design, now and after general availability: an administrator decides who can use it and, for each user, chooses whether the assistant can only view the data or also create and modify it. Even when enabled, the assistant operates with the account and permissions of the user behind it: whoever can't see the financials won't see them through AI either. MIS sends your data to no model on its own; when you connect your assistant, the terms of your AI subscription apply.
Getting connected
Four steps and you're in.
On Claude and ChatGPT the connector uses your MIS login: no keys to copy, no technical setup. On Gemini it runs through the CLI with a free API key — details in the requirements below.
1. In your AI assistant, add the Made In Show connector as a custom connector, with the URL provided by your administrator.
2. When the sign-in page opens, log in with your MIS credentials and approve the consent screen.
3. Your administrator enables AI access for your account: view, act, or nothing — decided per user.
4. Ask away. From then on the assistant works with your data and your permissions.
Requirements
All you need is the right AI plan.
The MIS connector plugs into the AI account each collaborator already uses. Here is what it takes, platform by platform.
ChatGPT (OpenAI)
A Business, Enterprise or Edu workspace is required: custom MCP connectors are enabled through developer mode by a workspace admin. Individual plans (Free, Go, Plus and Pro) currently do not support custom connectors.
ChatGPT plans and pricingClaude (Anthropic)
Works on every plan, including the free one (which is limited to a single custom connector — it can be MIS). On Team and Enterprise plans the organisation admin enables the connector, then each user links it to their own account.
Claude plans and pricingGemini (Google)
Works through Gemini CLI — Google's official free terminal tool — with a free Gemini API key, created in a minute on Google AI Studio. Paid plans only raise usage limits.
Gemini plans and pricing
Requirements verified on 25 August 2026 against the providers' official documentation. Plans change: we re-check them before every release.
Learn more: model training
The data the assistant reads from MIS flows through the user's own AI account and is processed under that provider's terms, as stated in the connector's privacy policy. MIS cannot see or control that account's settings.
On business plans (ChatGPT Business/Enterprise/Edu, Claude Team/Enterprise, Gemini with a Google Workspace or Gemini Enterprise account) providers exclude conversations from model training by default. On individual plans, use for training depends on an account setting — in some cases on by default — which each user can turn off in their provider's settings.
Status as of August 25, 2026: the MCP connector is in testing with our first pilot customers. It isn't generally available yet.
Want to be among the first?
Testing runs with a few customers at a time, by choice. If you want AI on your data before everyone else, book a demo and let's talk about your case.