The September reset: building the planning deck without losing your weekend

Second week of September. Three decks are running in parallel: the annual plan for leadership, the kickoff for the team, the budget trajectory for finance, who wanted it yesterday. Two weeks ago, everyone in the building believed this would be a calm month.

This is not bad luck. It is structural. Planning for the next fiscal year has to be framed while the current one still has a quarter left to run, which puts the thinking work in September and the number work in October and November. Whatever the calendar says about the working year, September opens planning season.

And planning season, in most organisations, is expressed in slides.

Why this particular deck is so expensive

We consistently underestimate what these presentations cost, because we count the cost in formatting hours when it is paid in attention.

Microsoft's Work Trend Index, based on a survey of 31,000 employees across 31 markets, gives the shape of the problem. Across Microsoft 365 apps, the average employee spends 57% of their time communicating (meetings, email, chat) and 43% creating. PowerPoint accounts for 8% of the total. Modest on paper, but that time is not spread evenly across the year. It clusters into a handful of peaks, and planning season is the tallest.

The more revealing number sits next to it: 68% of people say they do not have enough uninterrupted focus time during the workday, and 64% say they struggle to find the time and energy to do their job. An annual planning deck demands exactly that kind of sustained concentration. You have to hold a trajectory in your head, make trade-offs, decide what gets cut. Assembling it in twenty-minute gaps between meetings guarantees a document that wanders.

Hence the weekend. People do not sacrifice Saturday because formatting is slow. They sacrifice it because it is the only block of time in which nobody interrupts them.

The costly mistake: opening PowerPoint first

The most common error, and the most expensive, is to start in the slide editor.

The moment a blank slide appears, your brain switches into layout mode. Does the title fit on one line? Does the chart work better on the left? Does this icon read as serious enough? These are legitimate questions. They are simply the last ones to answer, not the first, and answered first they consume the one resource you are short of.

The order that works runs the other way.

1. Frame it in plain text. Before any slide, write one page describing what the deck must achieve, not what it must contain. "Get three budget trade-offs approved" is a different presentation from "align the team on the roadmap". Conflating the two is the leading cause of forty-slide decks: you try to do both, and you do neither.

2. Gather the material before you write. Last quarter's numbers, the framing notes, last year's slides. Collect first, filter second. The worst moment to go hunting for a figure is while drafting the slide that depends on it.

3. Decide the structure flat. One line per slide, in a plain text file. If the structure does not stand up as text, it will not stand up as slides. Design never rescues weak reasoning; it hides the flaw just well enough for it to surface in the room.

4. Format last. Ideally, not by hand.

What AI contributes here, and what it will not

Step four is where generation tools change the economics, and the boundary deserves a precise description.

What AI absorbs well: turning an approved structure into a built deck, applying the corporate template, producing layout variants, regenerating cleanly when a number moves. Everything mechanical and repetitive in the manufacturing.

What it will not do for you: decide. Microsoft's 2026 Work Trend Index, surveying 20,000 employees across ten countries, suggests users have internalised this. 85% treat AI output as a starting point rather than a final answer. That is the correct calibration. A tool that handed you a "finished" annual plan without your having arbitrated anything would mostly hand you a document you could not defend under questioning.

So the useful question is less "can AI build my deck?" than "what time does it give back, and am I reinvesting that time in the reasoning?" The same edition reports that 49% of AI users say they now produce work they could not have produced a year earlier. The gain is real. It rarely sits where people expect it.

The specific trap of the planning deck: the template

There is a concrete reason September decks go wrong more often than others: leadership looks at them.

A slightly ragged team deck passes. An annual plan presented to the executive committee with three different typefaces, a stretched logo and a footer missing from half the slides does not. This is where consumer-grade presentation generators break down: they produce attractive slides in their aesthetic, not yours. You get back a deck that has to be rebuilt from scratch, and the time you thought you had saved is gone.

That is the problem Paul treats as its first-order concern: applying your corporate typefaces, palette, masters and institutional footer rather than imposing a house style. For the most heavily codified formats (executive committee, board, steering committee), the board presentation use case sets out the expected structure and what editable PPTX export leaves you free to change afterwards.

The annual plan and the kickoff are not the same deck

One last confusion, and it costs a full weekend every time: the annual plan and the team kickoff share neither audience, nor objective, nor density.

The annual plan addresses people who must arbitrate. It is dense, quantified, and its quality is measured by how many decisions it lets the room take. One slide, one decision. Appendices sit apart from the main flow rather than interleaved with it.

The kickoff addresses people who must commit. It is more narrative, more visual, and its quality is measured by what the team still remembers three weeks later. A dense slide works against it.

Trying to serve both with one document produces a forty-slide deck that nobody reads end to end. The working practice is to build the annual plan first, then derive the kickoff from it by cutting, never by adding. It is also the case where regenerating from a shared base saves the most time.

The September checklist

Before opening anything:

• I have written, in one sentence, what this deck must achieve, not what it must contain.
• I know who decides in the room, and what they are being asked to approve.
• I gathered my numbers before I started drafting.
• My structure stands up as plain text, one line per slide.
• I have marked the slides that will also serve the kickoff.
• The corporate template is this year's, not the one from three years ago.

The rest (layout, visual consistency, variants) can be delegated. The reasoning cannot. That is good news, of a kind: the reasoning is the only part that earns you a seat in the room.

Sources

• Microsoft, Work Trend Index Annual Report, "Will AI Fix Work?": time split across Microsoft 365, focus time (survey of 31,000 employees across 31 markets, February to March 2023).
• Microsoft, Work Trend Index 2026: perception and use of AI at work (survey of 20,000 employees across 10 countries).