The Microsoft Teams SDK now lets developers port their existing AI agents into Teams without rebuilding anything from scratch. Using the HTTP server adapter pattern, your LangChain chains, Azure Foundry deployments, or Slack bots connect directly to Teams' messaging infrastructure. This means enterprise teams can access your agent where decisions actually get made.

ExpressAdapter wraps your existing Express server and injects POST /api/messages—the route Teams uses to deliver messages. The SDK handles request verification automatically and routes messages to your event handlers. Your server stays unchanged; the Teams SDK simply adds that one endpoint. The same three-step pattern works with the Python SDK using FastAPI.

Developers can run Slack and Teams bots side-by-side on the same Express process, share agent logic between both platforms, or connect LangChain chains with just a few lines of code. For Azure AI Foundry agents, the SDK forwards messages to your deployed agent and relays replies back to Teams. A typing indicator fires before LLM responses so users know something's happening.

Setup requires three steps: expose a public URL using Dev tunnels or ngrok, register the bot via the Teams CLI which handles AAD app registration and secret generation automatically, then sideload the app into Teams.