Microsoft has begun rolling out Autopilot, a new personal artificial intelligence agent designed to manage work, coordinate tasks and follow up on commitments without requiring users to provide instructions every time. Announced at Microsoft Ignite 2026 on November 17, Autopilot is being introduced through Microsoft’s Frontier early-access program as part of a broader transformation of the Copilot experience.
Unlike traditional AI assistants that mainly answer questions, summarise documents or generate content, Autopilot is designed to operate continuously in the background. Users can give it a name, assign it a role and set a goal. It can then monitor Teams channels, follow up on email conversations, coordinate meetings, manage recurring work and continue projects across multiple days. The agent has its own identity, memory, workspace and cloud-based computing environment, allowing it to keep working even when the user is offline.
Key Takeaways
- Microsoft announced the Autopilot personal agent at Ignite 2026.
- It is an always-on AI agent integrated into Microsoft Copilot.
- Autopilot works across Teams, Outlook, emails, meetings, chats and documents.
- Users can give it a name, role and specific objectives.
- It has its own identity, persistent memory and cloud-based computer.
- It can monitor conversations, prepare replies, schedule meetings and manage recurring tasks.
- It can coordinate complex workflows, including supplier reviews and project management.
- Microsoft is rolling it out to eligible Frontier customers.
- Access requires a Microsoft 365 Copilot licence, usage-based billing and appropriate administrative controls.
- Microsoft has not yet announced final general-availability pricing.
What Is Microsoft Autopilot?
Microsoft Autopilot is a persistent personal AI agent built into the redesigned Microsoft Copilot experience.
The company describes it as a digital teammate that can take responsibility for ongoing work rather than waiting for users to issue individual instructions.
For example, a conventional AI assistant might help write an email after receiving a prompt. Autopilot is designed to go further: it can monitor an email conversation, identify an unanswered question, prepare a response and follow up when necessary.
The agent can also maintain context across different projects and conversations. This means users do not have to explain the same background information every time they ask it to continue a task.
Microsoft originally introduced the technology under the name Scout in June 2026. It was subsequently renamed Autopilot as Microsoft expanded its vision for personal and enterprise AI agents.
The latest rollout brings the technology into the redesigned Copilot experience.
How Does Microsoft Autopilot Work?
Autopilot is designed around continuous task execution.
Instead of treating every interaction as a separate conversation, it maintains context about the user’s responsibilities, priorities and ongoing work.
1. Give Your Agent a Name, Role and Goal
Users begin by creating their personal Autopilot inside Microsoft Copilot.
They can assign it a name, describe its role and provide an objective.
For example, an employee could create an agent responsible for tracking a product launch.
The instructions might include:
- Monitor important project discussions.
- Track outstanding tasks and deadlines.
- Identify delayed decisions.
- Prepare weekly progress reports.
- Follow up with team members.
- Keep the project schedule updated.
Once configured, Autopilot can continue working towards the objective instead of waiting for a new instruction every day.
2. Autopilot Understands Your Work Context
One of the central technologies behind Autopilot is Microsoft’s Work IQ.
Work IQ provides business context by connecting information across Microsoft 365 applications, including emails, meetings, Teams conversations, files and organisational relationships.
This allows the agent to understand how work happens inside an organisation.
For example, if an employee asks Autopilot to prepare for a meeting, it could use relevant email conversations, previous meeting discussions and available documents to build the necessary background.
Microsoft says the agent works within the permissions already granted to the user.
This is important because enterprise AI systems must respect existing access controls rather than automatically gaining access to every document and conversation in a company.
3. It Continues Working in the Background
Autopilot is cloud-hosted, meaning it does not need the user’s computer to remain switched on.
Once assigned a task, it can continue monitoring relevant information and carrying out permitted activities in the background.
It can:
- Monitor selected Teams channels.
- Track email conversations.
- Follow up on unanswered requests.
- Run recurring tasks.
- Prepare meeting materials.
- Identify upcoming deadlines.
- Continue projects across multiple days.
This is one of the biggest differences between Autopilot and conventional AI chatbots.
The objective is not simply to respond when someone asks a question. It is to keep work moving even when the user is concentrating on something else.
What Can Microsoft Autopilot Actually Do?
Microsoft has highlighted several examples of how Autopilot could be used in everyday business operations.
1. Manage Supplier Reviews
One of Microsoft’s examples involves managing an entire supplier review process.
A user could ask Autopilot to organise the review, after which the agent could:
- Build a schedule.
- Prepare a workback plan.
- Identify important milestones.
- Coordinate meetings.
- Prepare supporting materials.
- Contact stakeholders for updates.
- Track outstanding responses.
- Follow up on unresolved issues.
This is a significant example because supplier reviews usually involve several departments and multiple rounds of communication.
Instead of manually coordinating every step, employees could delegate much of the administrative work to an AI agent.
2. Monitor Teams Conversations
Autopilot can work across Microsoft Teams conversations, channels and group chats.
Users can bring the agent into relevant discussions using mentions, allowing it to receive context and help coordinate follow-up activities.
For example, if a project discussion produces several outstanding tasks, Autopilot could help identify the responsibilities, track progress and remind participants about pending work.
3. Manage Emails and Follow-ups
Email management is another important use case.
Autopilot can monitor conversations, identify outstanding commitments and prepare responses.
It is designed to understand which messages require attention and which tasks need follow-up.
This could be particularly useful for managers, sales teams and professionals who handle large volumes of email.
However, consequential actions remain subject to organisational permissions and approval requirements.
4. Run Recurring Workflows
Users can also assign recurring responsibilities to Autopilot.
For example, an employee could ask it to prepare a weekly summary of project developments and deliver it on a specified schedule.
Other possible applications include:
- Weekly sales updates.
- Meeting preparation.
- Project progress reports.
- Outstanding task summaries.
- Follow-up reminders.
- Supplier status tracking.
The agent is designed to maintain the context required to repeat these activities without requiring users to recreate the instructions every time.
Autopilot Has Its Own Identity, Memory and Computer
One of the most important technical aspects of Microsoft’s announcement is that Autopilot does not simply operate as an extension of the user’s account.
Each agent has its own identity and workspace.
| Component | Purpose |
|---|---|
| Agent identity | Gives Autopilot a distinct identity within the organisation |
| Persistent memory | Retains relevant context across conversations and projects |
| Cloud computer | Allows work to continue when the user is offline |
| Workspace | Provides an environment for managing ongoing activities |
| Permissions | Restricts access to approved information and actions |
| Audit records | Helps organisations monitor what the agent does |
Microsoft says each Autopilot receives its own Microsoft Entra Agent ID and has a named manager.
This helps organisations understand which activities were performed by the agent and who is responsible for overseeing it.
The separation between a user’s identity and the agent’s identity is particularly important for enterprise security.
If an AI agent sends an email, accesses a document or performs an approved action, the organisation needs to know that the action was performed by an identifiable agent rather than an anonymous automated process.
Microsoft Introduces Security and Governance Controls
Giving an AI agent the ability to interact with company emails, files and meetings creates significant security responsibilities.
Microsoft says Autopilot is built on the identity, security and compliance systems already used by Microsoft 365 organisations.
Key safeguards include:
- Microsoft Entra Agent ID: Gives each agent a separate organisational identity.
- Scoped permissions: Limits which resources and systems the agent can access.
- Audit trails: Records the agent’s sign-ins and actions.
- Human approval: Allows important actions to require user confirmation.
- Administrative controls: Lets IT teams manage or block an agent.
- Work IQ permissions: Keeps the agent within the information access boundaries already established for the user.
These controls are intended to make autonomous AI agents more suitable for business environments.
Microsoft is also expanding Agent 365, its platform for managing, monitoring and governing AI agents across organisations.
The company is positioning governance as a central requirement for wider adoption rather than an optional feature.
Microsoft Autopilot vs Copilot Chat and Cowork
Microsoft is introducing several different experiences within its redesigned Copilot application.
Although they are connected, each is intended for a different type of work.
| Feature | Main purpose | Typical use |
|---|---|---|
| Copilot Chat | Answer questions and generate content | Summarising documents and drafting emails |
| Cowork | Complete a specific multi-step task | Preparing a report or coordinating a defined workflow |
| Autopilot | Manage ongoing objectives and recurring work | Monitoring projects and following up on commitments |
| Copilot Code | Build software and applications | Creating apps, dashboards and solutions |
The distinction is particularly important.
Cowork is designed for tasks that users hand over and expect to be completed. Autopilot is designed for responsibilities that continue beyond one request.
For example, Cowork could prepare a supplier comparison report. Autopilot could monitor the supplier review process over several weeks, track responses and follow up on outstanding information.
This positions Autopilot as a more persistent and proactive form of AI assistance.
Microsoft Autopilot Availability and Pricing
Microsoft began rolling out Autopilot to eligible customers through its Frontier program on November 17, 2026.
Frontier is Microsoft’s early-access programme for experimental and emerging AI capabilities.
The initial rollout is aimed at organisations with the necessary Microsoft 365 Copilot licensing and administrative setup.
| Detail | Current status |
|---|---|
| Product | Microsoft Autopilot |
| Platform | Microsoft Copilot |
| Initial rollout | November 17, 2026 |
| Access programme | Frontier |
| Microsoft 365 Copilot licence | Required |
| Billing | Usage-based |
| Administrative controls | Required |
| General availability | Not announced |
| Final pricing | Not announced |
Microsoft says Autopilot uses usage-based billing through Copilot Credits.
This means costs will depend on how much work the agent performs rather than being covered entirely by a conventional fixed subscription.
Administrators can establish spending policies and monitor usage to help control costs.
Microsoft has not yet published final pricing and packaging for general availability.
Why Microsoft Autopilot Matters for Businesses
The launch reflects a major shift in how companies are approaching AI adoption.
Until recently, many workplace AI tools focused on helping employees complete individual tasks more quickly.
Autopilot represents a different approach: assigning an ongoing responsibility to an AI agent.
Reduced Administrative Work
Employees spend considerable time coordinating meetings, following up on emails, tracking deadlines and preparing routine reports.
Autopilot could automate parts of this administrative workload.
Better Project Coordination
An agent that monitors discussions and outstanding commitments could help teams identify delays before they become major problems.
More Continuous Productivity
Because Autopilot operates in the cloud, work can continue outside normal working hours without requiring employees to remain connected.
More Personalised AI Assistance
Persistent memory and organisational context could help the agent understand an employee’s responsibilities and working preferences over time.
For businesses, the potential benefit is not simply faster content generation. It is reducing the coordination work that slows down projects and everyday operations.
How Microsoft Is Competing in the AI Agent Market
Microsoft’s Autopilot announcement comes as technology companies increasingly compete to build AI agents capable of performing real work.
OpenAI, Anthropic, Google and other AI companies are expanding their products beyond conventional chat interfaces.
The competition is moving towards systems that can use tools, access information, execute workflows and continue working across multiple applications.
Microsoft has a potential distribution advantage because Autopilot is integrated into the Microsoft 365 environment, which many businesses already use for email, meetings, documents and collaboration.
The company’s strategy is to combine AI capabilities with its existing enterprise identity, security and productivity infrastructure.
However, the success of Autopilot will depend on its reliability, cost and ability to complete complex workflows without creating additional administrative problems.
The Bigger Picture
Microsoft Autopilot highlights the transition from conversational AI to persistent workplace agents. Instead of asking an assistant to summarise an email or prepare a document, employees may increasingly delegate entire responsibilities that continue across days and weeks.
This could change how organisations distribute work between employees and software. Routine coordination, reminders, research and follow-ups are potential areas where AI agents could reduce manual effort. Microsoft’s integration with Teams, Outlook and Microsoft 365 gives it a direct route into existing business workflows.
However, autonomous agents also introduce new questions about accountability, security, data access and operating costs. Businesses will need to establish clear boundaries and monitoring systems before allowing AI agents to handle important work independently.
Looking Ahead
The immediate focus will be on Microsoft’s Frontier rollout and how organisations use Autopilot in real business environments. Companies will need to evaluate whether the agent can reliably manage ongoing projects, coordinate stakeholders and complete recurring tasks while respecting organisational permissions and approval requirements.
If Microsoft can make Autopilot reliable and affordable at scale, it could become a significant part of the next generation of workplace software. The broader opportunity is a shift towards AI systems that do not merely help people work faster but take responsibility for keeping important work moving forward.
Get the day’s top stories in your inbox
One concise email. No spam, unsubscribe anytime.



