Agentic Users are AI agents that join Microsoft Teams under their own identity, with their own Microsoft Entra account, an accountable human owner and admin controls, so they can be added to chats, channels and meetings like a colleague. Microsoft announced them at Microsoft Ignite 2026 (17–20 November, San Francisco and online), and they are now in public preview.
What did Microsoft announce? According to the Microsoft 365 Developer Blog (17 November 2026), an agent that acts as itself can now:
Microsoft also introduced Autopilot, a proactive personal agent with its own identity, memory and tools. Colleagues can @mention a person's Autopilot in Teams channels and group chats. Autopilot is rolling out through Microsoft's Frontier early-access program, and its work consumes Copilot Credits.
Why does it matter? For the first time, the agents your business uses will sit in the same chats, channels and meetings as your employees and alongside your Microsoft Teams telephony environment. Microsoft's own number shows the scale: organizations have registered more than 50 million agents in Microsoft Agent 365 since it launched in May 2026. Agents are no longer a pilot project. They are becoming a new class of user that IT has to provision, secure and audit.
What is the main risk of AI agents in Teams? Not the agents themselves, but agents that nobody owns, nobody watches and nobody can switch off. Three problems stand out:
Why is this urgent now? Microsoft is also expanding how agents behave inside collaboration tools: streaming responses in group chats and channels is available now, while threaded replies, visible plans and steps, and HTML widgets are in public preview. Each step makes agents more useful and more present in day-to-day conversations. That is exactly why a 24/7 security operations center needs to see agent activity alongside human sign-ins, endpoint alerts and email threats.
How should an enterprise prepare for agents in Teams? Treat every agent like a new employee with privileged access. A practical rollout looks like this:
What licenses do you need? Microsoft says core usage insights for Microsoft agents are now available with qualifying Microsoft 365 subscriptions, while advanced governance, security and insights for non-Microsoft agents require an Agent 365 or ME7 license. A managed IT services partner can map those entitlements to your current agreement before you commit.

Why should enterprises adopt agents in Teams at all? Because governed agents remove repetitive work from the places where teams already collaborate. When an agent has a clear identity, owner and scope, the business gets the productivity without the blind spots.
What else came out of Ignite for Teams? Microsoft also announced Teams Channel Backup in private preview, which restores channels and deleted or edited channel messages within a 30-day window, a welcome addition after the global Outlook and Teams outage that Satya Nadella addressed in his keynote.
How can HIT Communications help? With more than 30 years of experience serving enterprises across Latin America, the United States and Europe, HIT Communications helps organizations adopt Microsoft's agent platform without losing control of their collaboration environment:
Ignite 2026 made one thing clear: AI agents are becoming teammates inside Microsoft Teams, with their own identities, owners and audit trails. Agentic Users, Autopilot and the expanded Agent 365 platform give organizations the tools to manage them, but the tools only help if IT puts an inventory, ownership model and monitoring in place before agents spread across chats and meetings.
Start this week: list the agents already in your tenant, decide who owns each one, and pilot Agentic Users with a small, well-scoped team.
Want expert help preparing Microsoft Teams for AI agents? Contact HIT Communications to assess your environment, design a governance plan and protect every identity, human or agent, with our 24/7 SOC.

Find out how we can transform your business. Talk to one of our experts now!
Get in touch