Skip to content

Have Autopilot watch your deliverables and flag risks

Stop discovering slipped deadlines and stalled decisions after they've already cost you. Hand Autopilot (formerly Microsoft Scout) a standing watch over your deliverables and let it block the time and raise the flags for you.

Stage: Autopilots · For: Manager, Champion, End user · Level: Intermediate · Time: 15 min

When to use this

The things that hurt most are the ones that go quiet — a deliverable that crept up on you, a decision that's been "pending" for two weeks and is now blocking three people. Catching those takes constant attention, which is exactly what you don't have. An always-on agent does. Once Autopilot is enabled for you (see Find out if you can get Autopilot), this is the job that turns it from a convenience into a safety net.

As always, Autopilot works on your behalf, with its own identity in your tenant (permissions, audit and governance behind it) and keeps you in the loop — it raises risks and protects time, you decide what to do about them.

What you'll need

  • Access to Autopilot (private preview) — or, today, the Frontier desktop app still documented on Learn as Microsoft Scout (needs Frontier turned on by your admin, an Intune policy + admin attestation, a Microsoft 365 Copilot license, and a GitHub Copilot Business or Enterprise license). See Find out if you can get Autopilot.
  • Autopilot connected to the surfaces your work lives on — Teams, Outlook, OneDrive, SharePoint, and your calendar — so it can see what's due and what's stalling.
  • A rough sense of your upcoming deliverables and the decisions you're waiting on, so you can tell Autopilot what to watch.

Try it now — the prompt

Autopilot shows up where you already work — Teams, Outlook, chats, channels and documents — and you can @mention it like a colleague. Give it the standing watch in plain language:

Keep a standing watch on my deliverables and decisions:
- Identify what's coming due for me and automatically block calendar time so I
  actually have room to do it.
- Watch for things that are stalling — decisions that have gone quiet, items
  waiting on someone — and flag them to me before they turn into blockers.
- Give me a short heads-up when you block time or raise a risk, with the why.
Don't reschedule anything with [my leadership / customers] without checking first.

Why this works: it names the two grounded capabilities exactly — identify deliverables and block calendar time, and spot risks like stalled decisions before they block you — and it sets the human-in-the-loop boundary on the actions that carry weight.

Step by step

  1. Reach Autopilot where you already work (Teams, Outlook, a chat or a channel) and confirm it can see the surfaces your work runs on — calendar, Outlook, Teams, OneDrive, SharePoint.
  2. Hand it the standing watch. Paste the prompt above, naming the deliverables and the decisions you most need protected. This is an ongoing job, not a one-time check.
  3. Let it block time for what's due. Autopilot identifies upcoming deliverables and automatically blocks calendar time so the work has somewhere to live instead of getting squeezed out.
  4. Watch for the risk flags. Autopilot spots things that are stalling — a decision that's gone quiet, an item waiting on someone — and raises them before they become blockers, while there's still time to act.
  5. Decide and direct. When Autopilot flags a risk, you decide the move — nudge the owner, escalate, or let it ride. Autopilot surfaces; you steer.
  6. Refine what counts as a risk. If it's too noisy or too quiet, say so ("only flag decisions older than 5 days"). Autopilot builds context over time, so your tuning compounds.

Screenshots

We deliberately don't ship screenshots that go stale — the Microsoft Copilot UI changes often. Follow the numbered steps above, which we keep current. Maintainers can regenerate fresh captures with the Playwright tool in tooling/screenshots/.

Make it better

  • Protect the deep-work blocks. Once Autopilot is blocking time for deliverables, defend those holds — the value evaporates if you let every block get double-booked.
  • Turn flags into a weekly rhythm. Skim Autopilot's risk flags at a set time each week so spotting a stalled decision becomes a habit, not a fire drill.
  • Run it alongside meeting coordination. Deliverables and meetings are the same week from two angles — pair this with Let Autopilot coordinate your meetings and prep so both run in the background.

Watch out for

  • A flag is a prompt, not a verdict. Autopilot surfaces what looks stalled; whether it actually is — and what to do — is your call. It's an early-warning system, not a decision-maker.
  • You own the actions it takes. Autopilot acts on your behalf with its own identity in your tenant, with permissions, audit and governance behind it — keep schedule changes that affect leadership or customers behind your explicit sign-off.
  • Early on, it's still learning your work. Autopilot builds context over time, so expect the first weeks of flags to need more tuning than later ones.
  • It's private preview. Capabilities can shift as Autopilot evolves — verify against the Microsoft Scout desktop app docs if behavior differs from what's described here.

Where this leads (the ramp)

You've now handed Autopilot two standing jobs and felt what an always-on agent does. When you hit the wall of what a ready-made Autopilot can do — you need custom logic, real actions against a system of record, or org-wide governance — that's the signal you've outgrown adopting an agent and you're ready to build one.

Next: Stage 6 · Copilot Studio — when you need to build custom actions, logic, and governance

Was this page helpful?

Found an issue or want to suggest an improvement? Open a GitHub issue  ·  Star on GitHub