Microsoft 365 Copilot Plan My Day Template - Building a Daily Briefing Agent

Every executive I've worked with has the same morning ritual. Coffee. Phone. Email. Then twenty minutes of trying to work out what actually matters today versus what's just noise filling up the inbox. The promise of an AI-generated daily briefing has been around for years, but the implementations have mostly been disappointing. They either summarise everything (useless) or summarise the wrong things (worse than useless).

Microsoft's Plan My Day template in Agent Builder is the first one I've used that gets close to being genuinely helpful out of the box. It still needs work for most organisations, but the starting point is solid enough that you can have something useful running in an afternoon. Let me walk through what it does, what's missing, and how we'd typically extend it for an Australian leadership team.

What Plan My Day Actually Does

The template ships as a declarative agent inside Microsoft 365 Copilot Agent Builder. Once enabled, it pulls signals from across your Microsoft 365 footprint and generates a structured morning briefing. We're talking calendar, Outlook, Teams messages, SharePoint activity, OneDrive changes, and even personal events from your calendar.

The output is opinionated, which I like. Most AI briefings dump everything on you and let you sort it out. Plan My Day actively ranks items by business impact, time sensitivity, who is blocked waiting on you, and strategic alignment. It surfaces the top three to five priorities. The rest goes underneath, scannable but not in your face.

The tone Microsoft has gone with is interesting. They've described it as a "chief of staff" tone. Warm, efficient, gives you what you need without padding. When I first tested it I expected the usual AI overproduction with three exclamation points per paragraph. Instead it was almost terse. Items get linked directly. Empty sections get omitted entirely. You can scan the top of the briefing in 30 seconds and read the full thing in 5 to 7 minutes. That ratio matters because executives are not going to read an essay every morning.

The Features That Actually Help

A few capabilities stood out when we tested this with a client's leadership team.

Multi-day awareness. The agent doesn't just tell you about today. It looks ahead to tomorrow and the next three working days. Meetings that need prep get flagged early. Deadlines approaching get a warning before they become urgent. External engagements get surfaced so you actually remember to think about them. The agent is also smart about weekends and public holidays, which is more impressive than it sounds. I've seen "smart" calendar assistants in the past that helpfully reminded executives about meetings on Australia Day.

Decision queue management. This is the one feature I've been waiting for. The agent identifies items where someone else is blocked waiting for your input. Approvals you haven't given. Replies people are chasing. Reviews that have been sitting in your inbox for three days. It orders these by how long the person has been waiting, longest first. Sounds simple. Try doing this manually every morning and you'll quickly see why it isn't.

Conflict detection and delegation suggestions. The agent flags overlapping meetings, double bookings, and overloaded days. The definition of "overloaded" is more than six meetings, which feels about right. It will then proactively suggest items you could delegate, decline, or shorten. This is the kind of thing a real chief of staff would do, and I genuinely didn't expect the AI to handle it as well as it does.

Personal wellness touches. This is where Microsoft has done something a bit different. The agent surfaces team birthdays, work anniversaries, and family events from your personal calendar. With reminders at one week, three days, and day of. This feels small until you've worked with an executive who keeps forgetting their direct report's anniversary. Then it feels significant. I'll admit I was sceptical of this feature when I first read about it. After two weeks of using it I changed my mind.

Where the Template Falls Short

This is where I have to be honest about what the template is and isn't.

The Plan My Day agent only knows about what's in your Microsoft 365 tenant. That sounds obvious until you remember that most executives' day is also shaped by what's happening in their CRM, their project tracking tool, their incident management system, and their financial dashboards. Out of the box, the agent has none of that context.

The other limitation is the one-question-at-a-time constraint. The template is designed to answer one prompt cleanly. If you ask it "give me my day and also let me know what's happening with the Westpac account and also tell me how the Sydney office is tracking on hiring," it gets confused and the response quality drops. Users need to be trained to ask one thing at a time, which is fine, but it isn't how most people naturally talk to AI.

There's also a disclaimer worth taking seriously. The agent uses generative AI. It will sometimes get things wrong. The default disclaimer reminds users to verify accuracy, especially for financial decisions. I've seen organisations remove this disclaimer because it felt unnecessary. I'd push back on that strongly. The cost of getting one priority wrong on a board meeting day is significantly higher than the mild friction of a disclaimer.

Extending It Properly

The template is a starting point. The real value comes from connecting it to the systems where the rest of your business actually runs. We typically work with clients to do four things.

First, connect a project tracking tool through a Microsoft 365 Copilot connector. Azure DevOps, Jira, ServiceNow, whichever one you use. Now the briefing includes open work items, sprint deadlines, and assigned tasks alongside the calendar and email signals. The priority ranking gets meaningfully better because the agent can see actual project commitments, not just whatever showed up in your inbox.

Second, add team OKR or quarterly goals as a SharePoint knowledge source. The agent's priority engine works better when it can weigh items against documented strategic objectives. If your top OKR for the quarter is winning two enterprise deals in Queensland, the agent will weight related calendar items and emails higher. The output starts to feel like it actually understands what you're trying to do, not just what's filling up your time.

Third, ground the agent in organisational context. A SharePoint site with the org chart, team rosters, and VIP contact lists. Now the agent knows that an email from the CFO is more important than an email from a vendor, even though both might be unread. It can personalise delegation suggestions because it knows who reports to whom.

Fourth, schedule the briefing. Most executives want their daily plan at 7am, sitting in their inbox or Teams chat ready to read with the morning coffee. Setting up a scheduled prompt to deliver this is straightforward and changes how the agent gets used. Without scheduling, the agent only runs when the executive remembers to ask. With scheduling, it becomes part of the morning routine.

We help clients design and deliver these extensions through our agentic automation work and through the Copilot training we run for leadership teams who want to use these tools properly rather than just dabble.

Who This Template Actually Suits

A few patterns from the rollouts we've done.

This template works brilliantly for executives whose work is largely driven by meetings, decisions, and people. Heads of operations, sales leaders, account managers, anyone with a packed calendar and a busy inbox. The signal-to-noise ratio improves visibly within a week.

It works less well for individual contributors. If your day is mostly focused work with a small number of meetings, the agent has less to work with and the briefing feels thin. You're better off using Copilot directly for the specific tasks you need.

It works very well for chiefs of staff who actually exist as a role. Sounds counterintuitive. The agent doesn't replace the human chief of staff. It does most of the data-gathering work that the human used to do, freeing them up for the harder stuff like reading the room, managing politics, and thinking strategically about what to bring forward.

Watch Outs for Australian Businesses

Two specific things to think about.

Compliance with the Privacy Act and any sector-specific data handling requirements. The agent processes employee data through its priority calculations. The work anniversary reminders are using HR data. The family event awareness is using personal calendar data. For most organisations this is fine because the data already lives in Microsoft 365 and the processing happens inside the tenant. But for regulated industries, run this through your privacy team before rollout. We've covered this in our broader AI strategy work for leaders, and the patterns transfer here.

Local labour laws and HR policy compliance. The agent may surface delegation suggestions or workload concerns that touch on employment matters. Make sure your managers know that AI-generated suggestions about workload, delegation, or capacity are starting points for a human conversation, not decisions in themselves.

The Bottom Line

Plan My Day is the strongest out-of-the-box agent template I've seen from Microsoft. It is genuinely useful with no configuration, and meaningfully more useful with about four hours of customisation. For Australian businesses already on Microsoft 365 Copilot, this is one of the lower-effort wins available right now.

The catch is the same as every other AI rollout. The technology works. The hard part is getting people to actually use it, and getting your data clean enough that the AI's recommendations are worth following. Plan My Day will surface the wrong priorities if your calendar is full of meetings nobody needs to be in and your emails are flagged inconsistently. The agent reflects your information hygiene back at you, in colour.

Reference: Plan My Day template documentation