There is no single “ChatGPT for Teams” app, so the best way to use ChatGPT in Microsoft Teams depends on which of four things you want: a bot in your channels, Microsoft’s own Copilot, ChatGPT reaching into Teams from the outside, or a workflow you build yourself. PlugAnd.ai is the turnkey option we found that puts OpenAI’s GPT models in a shared channel without per-user seats.

Most roundups blur these together, describing Microsoft’s Copilot, OpenAI’s Teams app and a channel bot alike as “ChatGPT in Teams”. They differ on who can see the answer, who pays for a seat and what data the assistant can read.

Four meanings of “ChatGPT in Teams”

An assistant in the channel. A colleague @-mentions a bot, it answers in the thread and everyone sees the reply. This is what most people picture. It requires a third-party app or something you build, and it runs GPT models rather than the ChatGPT product.

Microsoft’s Copilot. Microsoft’s Learn overview describes a free-with-eligible-license Copilot Chat grounded in the web and a paid Copilot (Premium) grounded in your tenant. You use Microsoft’s interface, and Microsoft’s page does not name the underlying models.

ChatGPT reaching into Teams. OpenAI’s Teams app searches chats and channels from inside ChatGPT. The work happens in ChatGPT, not in Teams.

A workflow. A message arrives, a scenario calls the OpenAI API, a reply is posted. Useful for one job, not a conversation.

Free is a real option, and it changes the ranking

A team may not need to buy anything. Per Microsoft Learn, Copilot Chat (Basic) comes with eligible Microsoft 365 licenses, appears in Teams and is covered by Enterprise Data Protection. If your questions are general (drafting, explaining, brainstorming), it may be enough. Pay for something else only when you need a shared channel assistant, a specific model, or answers from your own data.

The model choice moves the bill

PlugAnd.ai’s pricing page includes a worked example for a ten-person tenant: 4,200 requests a month at 3,000 input and 700 output tokens each. On GPT-5.6 Luna ($0.20 input and $1.20 output per million tokens) that is about $6.05 of usage, or about $6.72 of top-up including the 10% fee; on Claude Sonnet 5.5 ($2.00 and $10.00) about $54.60, or $60.67 of top-up. We re-ran the arithmetic and it holds. The vendor labels it an illustration; long threads, attachments, reasoning tokens and web search all add cost.

On a usage-billed assistant, use the stronger model deliberately, for the questions that need it. A per-seat license hides that trade-off; a shared balance shows it.

Who should pick what

A team that wants to ask GPT in channels, with no seats: PlugAnd.ai, using a cheaper model for routine questions.

A team that wants private, general answers at no extra seat cost: Microsoft Copilot Chat (Basic), if your licenses include it.

A company that needs answers from its own chats, files and meetings: Microsoft 365 Copilot (Premium).

People who already pay for ChatGPT and want it to search Teams: OpenAI’s Teams app, after your admins approve the permissions.

An IT team building governed, domain-specific assistants: Copilot Studio agents, budgeted on credits.