Microsoft has officially introduced Microsoft Scout, its first-ever “Autopilot” AI agent, a persistent, always-on autonomous assistant designed to operate continuously across Microsoft 365 apps without waiting to be prompted.
Unveiled at Microsoft Build 2026 on June 2, Scout represents a fundamental shift in how AI integrates into enterprise workflows.
Microsoft is launching a new category of AI agents it calls Autopilots, always-on, identity-bearing systems that act autonomously on a user’s behalf.
Unlike traditional AI assistants that respond only when queried, Autopilots remain active in the background, monitoring signals, reacting to triggers, and resuming tasks without user initiation. Scout is the first of this new class, and Microsoft has indicated more Autopilots will follow.
Microsoft Unveils Scout AI Agent
Scout integrates natively across the full Microsoft 365 ecosystem, connecting to Teams, Outlook, OneDrive, SharePoint, and underlying data layers including chats, email, calendar, and contacts.
Users interact with it primarily through Teams, while the desktop app extends Scout’s reach into the browser, local resources, and Model Context Protocol (MCP) servers.
The agent is powered by OpenClaw open-source technology and Microsoft’s proprietary Work IQ layer. Work IQ enables Scout to build persistent context over time, learning how a user works, what their priorities are, and which tasks require forward action.
This accumulated understanding makes Scout progressively more relevant and actionable as it operates within a specific user’s environment.
Key capabilities Scout performs autonomously include:
- Scheduling and coordinating meetings across time zones
- Flagging high-priority meetings and generating preparatory materials
- Blocking calendar time for identified deliverables
- Surfacing risks such as stalled decisions before they escalate into blockers
From a security perspective, Microsoft Scout is architected with enterprise-grade identity and access controls that distinguish it from consumer AI tools.
Every Scout instance operates under its own governed Microsoft Entra identity, not a shared or anonymous service account, ensuring all agent actions are attributable to a known, directory-managed actor.
Credential protection is enforced end-to-end: credentials are scoped to the task at hand, redacted from logs and diagnostics, and managed under Microsoft’s first-party service standards.
Sensitive actions require human approval before execution, and Microsoft Purview policies, including sensitivity labels and data loss prevention (DLP) rules, are enforced in real time before any data is written or transmitted.
Scout also contributes back to the open-source community by pushing policy conformance capabilities upstream to OpenClaw, enabling organizations running OpenClaw independently to validate whether their environments meet security and compliance requirements with audit-ready verification.
Microsoft Scout is currently in private preview, initially available to a select group of Frontier program customers and Microsoft employees who have already been piloting the experience internally. Access requires:
- Frontier program enrollment
- Intune policy configuration
- An opt-in attestation
- A valid GitHub Copilot Business or Enterprise license
Full setup documentation is available via Microsoft Learn. Broader availability details have not yet been announced, though Microsoft has confirmed all Microsoft 365 subscribers will eventually be eligible for Frontier access.
Free Webinar on OWASP API Top 10 and Guide to Close Visibility Gaps With WAAP