April 22, 2026
Agents assemble, users grumble
Bring your own Agent to MS Teams
Plug in your AI buddy to Teams; commenters want fixes, not fanfare
TLDR: Microsoft made it easy to plug existing AI bots into Teams with one simple server hookup, saving devs time. Commenters clapped back, saying fix basics first—Mac screen-share lag, message delivery issues, and a missing command-line tool—while reigniting the eternal Slack vs. Teams flame war.
Microsoft’s new “bring your own agent” trick promises you can drop your existing AI helper—whether it’s a Slack bot, a LangChain setup, or an Azure AI Foundry agent—straight into Teams with a simple adapter. In plain speak: add one inbox route to your server, and Teams starts talking to your bot while the toolkit handles verification and routing. Fewer rewrites, one codebase, faster rollout—that’s the pitch.
But the comments? Spicy. Instead of applause, the top asks are “please fix the lag,” especially on Mac screen sharing, and “please make messages actually deliver.” One user roasts the app for “silent failures… in Year of our Lord 2026,” while another demands a simple developer tool: “Why no Teams CLI?” The vibe swings from practical (save time by running Slack and Teams in the same app) to theatrical, as one commenter brands living in Teams chat as “hell” and pledges loyalty to Slack—even with GIFs censored. Meanwhile, the “agent about agents building agents” meme makes an appearance, poking fun at the AI hype spiral.
So yes, it’s a clever shortcut: one server, two chat worlds, less duplicate work. But the crowd’s chorus is clear: great feature, wrong order. Before new toys, they want boring-but-essential fixes, better tooling, and fewer “where did my message go” mysteries.
Key Points
- •The Teams TypeScript SDK uses an HTTP server adapter to integrate existing Express servers with Teams.
- •Initializing the SDK injects a standard POST /api/messages endpoint used by Teams to deliver bot messages.
- •The SDK verifies requests originate from Teams and routes messages to appropriate event handlers.
- •Scenario 1 shows co-hosting Slack (via Bolt/ExpressReceiver) and Teams bots on the same Express process, sharing business logic.
- •Scenario 2 bridges Teams messages to a LangChain chain with a typing indicator; Scenario 3 forwards messages to an Azure AI Foundry agent.