The board deck slide we keep having to write.

Every AI-strategy conversation at board level, in every organisation we’ve walked into, seems to end up needing the same slide. It’s usually slide two or three, and it’s usually the slide the rest of the deck should have been ladder-building to.

The slide is simple. Two axes, four quadrants, a small population of labelled dots.

The slide

                 Would it work if we built it?
                 ↑
       YES       │
                 │
   Strategic     │   In-flight
   priorities    │   programmes
                 │
     ────────────┼────────────→   Would it move a P&L line?
                 │                 (Yes →)
   Wishful       │
   thinking      │   Cost-only
                 │   automations
                 │
       NO

Every candidate initiative goes on the plot as a dot. Every dot gets a label. The labels are the names of the initiatives the audience already knows about, from other slides in other decks. The plot is where those initiatives actually sit.

Why this slide keeps working

The slide works because it forces two conversations the audience has been having separately.

“Would this move a P&L line?” is a business question. It’s the question the board should be asking. Most AI initiatives sit on the right of the plot without meaning to; they’re being justified by cost savings that don’t survive contact with the operating model.

“Would it work if we built it?” is a technical question. It’s the question the CTO should be asking. Most AI initiatives sit above the line because the technology can do the demo; whether the demo would survive in production is a different question, and it’s the one the technical team is best placed to answer.

Overlay the two, and the initiatives cluster.

In-flight programmes cluster in the top-right — legitimately. If they were shipped, they’d move a number. If they were built, they’d work. These are the ones to keep funding.

Strategic priorities cluster in the top-left. Technology is convincing; commercial case is thin. Usually because the initiative is solving a problem the business doesn’t have, or the P&L line it would move is a rounding error.

Cost-only automations cluster in the bottom-right. Business case is solid; the technology to deliver it doesn’t work at the reliability the business needs. Usually because “reliability” got defined at demo-standard, not at operational-standard.

Wishful thinking clusters in the bottom-left. Neither test passes. These are the initiatives on the deck because somebody added them at a different meeting.

What the slide changes

Two things, in every conversation we’ve watched.

Initiatives move. Almost every deck we’ve seen puts every initiative in the top-right, because the slide before was a slide of business cases. When the initiatives get relocated onto the two-axis plot with the board in the room, three or four move. That’s the useful part of the conversation.

Fewer initiatives get funded. The board is much better at cutting initiatives it can see than at cutting initiatives it can’t. Two axes, one plot, dots you can label: it’s the affordance the board didn’t have in the previous deck.

The version that doesn’t work

Three ways this slide fails:

Too many dots. More than ten and the plot is decoration. Cluster the initiatives before you plot them; put the clusters on the axes.

Axes that don’t force a decision. “Impact” vs “Effort” is a lazy version of the same plot that lets every initiative sit in “high impact, low effort” because everyone’s optimistic. The specific axes above — commercial and technical — are the ones that force honest dots.

The audience getting to place the dots themselves. You want the audience to react to the placement, not to decide the placement. Ideally, you’ve done the plot in a working session with the sponsor before the board sees it. What lands on the board slide is a stable picture, not a debate in progress.

Why it belongs here

The slide is not clever. It doesn’t require the audience to learn a new framework. It doesn’t need the analyst who made it to explain what the quadrants mean. It’s just: “here’s what we’ve been calling our AI strategy, plotted honestly against what would work and what would matter.”

Every time we’ve drawn it, the audience response has been the same short pause followed by the same short sentence: “so we should stop doing these ones on the left, then”. Which was, most of the time, the whole point of the deck.

So what?

If you’re being asked to present an AI strategy at board level, and the audience is smart and skeptical, this is the slide that reliably makes the rest of the deck coherent.

If you’re being asked to present an AI strategy and the audience is neither smart nor skeptical, this is still the slide that makes the strategy survive the year that follows.

Related reading

• What “AI transformation” actually costs the second year — the operating budget every top-right dot has to survive.
• The AI Opportunity Sprint — the engagement shape this slide falls out of.