The People AI Needed Were the Ones We Laid Off.

The People AI Needed Were the Ones We Laid Off.

In May, Gameometry made a simple argument: the hardest part of adopting AI in games isn't the AI. It's the operating system: the cadences, decisions, and measurement infrastructure AI is supposed to multiply. When pilots don't compound across a portfolio, the broken piece is almost never the tool. The research and advisory firm Forrester titled its June read on agentic AI like a verdict: companies are chasing, few are catching. Three-quarters of enterprise leaders say they're adopting agents. Only a sliver have them doing real production work.

The diagnosis is public. Studios have the data. So why is the gap widening instead of closing?

Best conversion ever

Two midcore titles, call them Game One and Game Two, ran the same tentpole event in the same month. Both recap decks claimed a 'new conversion high-water mark!' and both claims were true. Three weeks later they were different games.

On Game One, the event offer had discounted hard enough to flip the fence-sitters, who spent once before going quiet. Baseline conversion re-settled below where it started. On Game Two, the event reactivated lapsed payers who kept paying after it ended. The right calls were opposites. For Game One, stretch the payment interval before deep discounting trains players to wait for the sale. For Game Two, run the event monthly and rotate the content. After the event, pull repeat-purchase cohorts on both games and confirm 'no harm done.'

That's operating knowledge. It reads what the dashboard leaves out and knows which question comes next. Years in the seat build it.

What the cuts removed

The senior operators who carried this judgment were often the ones let go.

And it isn't past tense. By TradingPlatforms' June count, the year had already logged more than 3,700 verified cuts before the Xbox news: Epic took more than 1,000 of them in March, a fifth of the company, and Ubisoft is on its sixth round of the year. On July 6, Microsoft announced an Xbox overhaul with 3,200 roles on the block, even as it budgets $190 billion for AI infrastructure in 2026. Almost nobody making those cuts believed AI would replace judgment. Epic and Microsoft both put it on the record: the cuts weren't about AI. The institutional knowledge AI needed to work got cut all the same.

Hand that seat to a generalist who learned the category last year and the gap widens. No amount of tooling makes up the difference.

Studios didn't just cut headcount. They cut the judgment that knew what to do with it.

McKinsey's State of AI survey from last November found only 6% of nearly 2,000 firms capture meaningful EBIT impact from AI. Those that do are nearly three times as likely to have fundamentally redesigned their workflows. The separator is the redesign underneath.

Much of today's speed pressure comes from boardrooms optimizing for earnings per employee rather than long-term operating capability. Under that math, AI exists to shrink the denominator, and the fastest way to shrink it is to cut the expensive people who hold the judgment. But speed was never the constraint on a slate worth shipping. Run the same uninspired slate through faster pipelines and you get the same mediocrity, with the misses arriving ahead of schedule.

The diagnostic is one question.

If your best operator left tomorrow, would decision quality hold for six months? If not, the judgment lived in a person, not a system. The layoffs already axed them.

For studios that want the judgment back, hire operators who've already run organizations where cross-portfolio judgment mattered. Knowing AI is the easy half. The judgment is much harder to imitate (sorry, Claude).

The next generation of winning studios will preserve and compound operating judgment while everyone else mistakes faster execution for better execution.

The technology is ready. The infrastructure to absorb it was restructured out.

AI is the multiplier. Judgment is what you're multiplying.

Gameometry is built for this work. AI visibility, readiness diagnostics, decision-cadence redesign, and the operating-system reshape that turns AI into actual margin live in the same operational frame. Reach out at https://gameometry.io/contact if your studio is sitting on any of them.