The AI Hiring Bet That's Already Going Wrong

Fortune published the numbers last week, and they're worth sitting with. Sixty-six percent of CEOs plan to freeze or cut hiring through the rest of 2026. Since 2025, corporate America has eliminated over 1.17 million positions to fund AI investments. Entry-level job listings have dropped 30% since 2022. Middle management postings are down 42%. The bet is clear: route the headcount budget into AI infrastructure, show better margins, and let the tools carry the output.

I've seen this logic up close. I've also seen what it costs.

I built design teams from scratch twice. At Metro Bank, we started with zero and grew to a twenty-five person design organisation over three years. At Almosafer, I inherited fourteen designers and scaled through a period of significant product expansion. In both cases, the hardest work wasn't hiring people. It was building the conditions where designers developed the kind of judgment that separates a product that works from one that just ships.

You can't buy that judgment when you need it. There's no recruiter who can fill the role of 'designer who has been in enough product failures to recognise what one looks like before it happens.' That capability compounds over years of shipping, reviewing, being wrong, and adjusting. When companies freeze their hiring pipeline in the name of AI efficiency, they're not just delaying headcount. They're halting the development of the one thing AI can't replicate.

The Fortune piece was specifically about middle management, the connective tissue roles that implement and scale strategy. Design leadership lives exactly in this category. Eighty-four percent of the CEOs surveyed acknowledged that real AI ROI takes multiple years. Fifty-three percent of investors expect returns in six months. That gap doesn't get filled by a tool. It gets filled by people who know how to make the tool useful.

This is where the design conversation gets specific. AI tools generate output fast. They compress the distance between 'we need a screen' and 'here is a screen.' What they don't compress is the distance between 'here is a screen' and 'here is the right screen.' Every expensive product failure I've been called in to fix has lived in that second gap. Onboarding flows that tank activation because the copy assumes financial literacy the user doesn't have. Checkout redesigns that drop conversion by 15% because friction was removed in the wrong place. AI-powered features that users find confusing because three people with different assumptions each approved a different part of it without a designer holding the thread.

These are judgment failures. And judgment is what you lose when you stop developing the people who carry it.

When I was leading the Krak neobank delivery, a seven-million-dollar product shipped in eight weeks, over thirty thousand card orders in the first weeks post-launch, we used AI across the design workflow. MCP-powered Figma, automated component generation, compressed handoff. What made it work was not the tooling. It was senior designers who could look at AI output and make fast, accurate calls about what was right and what needed to change. That combination is what scales. The tools without the judgment is just fast.

Running RUONALIM in Dubai, I see both sides of this. Companies investing in senior design judgment are shipping better products. Companies that cut it are coming back six months later asking agencies like ours to fix the decisions nobody was qualified to make.

If you're a design leader, this is the conversation to have before the freeze, not after. The argument isn't 'protect design headcount.' The argument is that the cost of removing senior design judgment from your product organisation is not zero. It just doesn't show up on this quarter's spreadsheet. Make them look at the right timeframe.

The bet on AI is not wrong. The assumption that it replaces human judgment is.

---