Under pressure to prove AI is paying off, most leaders reach for the easiest available metric: fewer new hires, a role that "didn't need replacing," a headcount line that's flat or shrinking. It reads well in a board deck. For most businesses, it's also the wrong number to be optimising for.

Here's the problem with headcount as the scoreboard. Framing AI success as "this made a role redundant" saves cost in the short term, but it tends to damage two things at once. Internally, remaining employees adopt AI more cautiously and defensively once they sense the narrative is "AI is here to replace people" — which slows exactly the kind of confident, skilled use leadership actually needs from them. Externally, thinned teams often end up using AI to patch gaps rather than to genuinely elevate the work, which shows up as quality strain customers eventually notice.

The companies actually capturing outsized value from AI mostly aren't the ones with the deepest headcount cuts. They're the ones where the team stayed roughly the same size and started producing measurably more — more accounts handled well, more output per person, faster turnaround, better judgment calls at scale. That's a harder story to reduce to a single line item, which is exactly why it gets under-reported.

The strategic mistake, stated plainly: treating "we didn't need to hire that role" as the win condition, instead of "our existing team now produces meaningfully more than it did before." Most businesses never track revenue or output per employee closely enough to even notice when this is the real story — so the only metric anyone reports is the one that's easiest to count, not the one that's actually true.

The reframe worth putting in front of the board: the number that matters isn't heads removed. It's output or revenue per employee, before and after. It's a harder number to fake, and a far more durable growth story to tell.

We frame every AI rollout around what a team can now do, not who it can now do without — because the first story holds up under scrutiny and the second one usually doesn't. If your business is trying to figure out what the honest ROI story actually is, reply and we'll help you find the real number.

A quick note: last week's Friday Wrap-Up didn't go out — a run of severe power outages in our area knocked things offline for longer than expected. Apologies for the gap, back on schedule this week.

The Tuesday Briefing is published weekly by The Xylora Digest.