From Copilot to Autopilots: What’s Actually New
Microsoft Autopilots agents are autonomous AI assistants built into Microsoft 365 that monitor everyday activity across apps and then take background actions, helping schedule, coordinate, and unblock routine work without waiting for users to prompt them each time. This marks a change from Copilot as a conversational aide to agents that act with their own identity inside a tenant. Announced at Build, Autopilots are described by Satya Nadella as “always-on agents that work autonomously, with their own identity, and act on your behalf.” Scout is the first example, embedded directly into Teams chats and Outlook threads as a full participant instead of a sidebar bot. The pitch is simple: keep work in motion when people are busy, away, or context-switching. That means handing over recurring tasks—like chasing decisions or preparing for meetings—to AI that runs continuously in the background.

Scout: An Always-On Agent Living Inside Teams and M365
Scout is Microsoft’s first Autopilot agent and sits across Teams, Outlook, OneDrive, SharePoint, and calendar, tied to work data such as chats, email, and contacts. Unlike earlier Copilot experiences, Scout joins Teams group chats, participates in Outlook email threads, and stays active across cloud, desktop, and web. Scout carries its own Entra ID so every action is linked to a traceable, policy-aware identity. According to Omar Shahine, “Microsoft Scout is integrated across the Microsoft 365 apps you use every day, keeping it grounded in your flow of work.” Users can talk to Scout in Teams like a colleague, but most of its value comes from always-on AI background tasks: watching deadlines, surfacing key messages, and coordinating updates without needing constant nudges. That shift turns AI in Microsoft 365 from a tool you consult into a semi-autonomous presence inside daily collaboration spaces.
How Autonomous AI Assistants Keep Work Moving
Microsoft is positioning Autopilots as AI agents for workplace automation that handle coordination and follow-up so people can focus on higher-value work. Scout can schedule meetings across time zones, block focused time before deadlines, and prepare materials or briefings ahead of important sessions. It also scans for stalled decisions, nudging participants or highlighting risks before they become full blockers. These autonomous AI assistants watch ongoing activity across Teams and M365 to understand “how work gets done across your apps and systems,” then act within the permissions and policies set by each organization. Instead of repeatedly asking Copilot to draft messages or check calendars, users let Scout run continuously. The result is an AI layer that behaves more like a diligent project coordinator: keeping threads alive, aligning calendars, and pushing tasks forward even when a user’s attention is elsewhere.
Work IQ and OpenClaw: The Tech Behind Scout’s Autonomy
Scout’s ability to operate with context comes from two main ingredients: Work IQ and OpenClaw. Work IQ is an intelligence layer in Microsoft 365 that reads signals from email, files, meetings, and calendars to infer priorities, relationships, and active projects. Charles Lamanna describes it as the layer that “allows agents to plan, act, and produce outcomes that are grounded in how your business runs.” Any agent built on the M365 stack will be able to call this layer through APIs. Under the hood, Scout runs on OpenClaw, an open-source autonomous agent framework that quickly gathered strong developer interest after launch. Instead of building a competing stack, Microsoft is adopting OpenClaw and contributing enterprise-grade policy and governance features back to the project. This combination gives Scout both organizational awareness and a flexible agent engine, positioning it as a template for future Microsoft Autopilots agents inside the wider M365 ecosystem.
Trust, Governance, and the Road to Truly Autonomous Agents
Handing everyday decisions to always-on agents raises clear questions about control and trust. Scout operates within Microsoft 365 tenant policies, and organizations can define what data Autopilots can access and what actions they can perform. Scout’s dedicated Entra ID helps with audit trails, but many IT teams are already asking for stronger tenant-wide controls, which Microsoft says are still in development. Users will be able to customize Autopilots with names, speaking styles, context, and memory, but the deeper change is cultural: people must accept AI agents that act first and inform later. Early access is rolling out through Copilot Frontier and broader previews, while Scout remains an add-on for M365 E3 and E5 customers. Together, Autopilots and Scout signal a shift from chat-based Copilot interactions toward truly autonomous AI agents that keep workplace processes running without constant human direction.






