· AI
Most 2026 AI Restructurings Failed a Simple Test: Would Anyone Outside Take the New Job
A restructuring that keeps the headcount only counts if the new jobs are real, and Meta's own numbers show how often they aren't.
Almost every large "AI restructuring" in 2026 has followed the same shape: layoffs, a smaller group moved into AI-labeled roles, and a memo calling the whole thing a reorganization rather than a cut. Monday.com cut about 620 people, roughly 20 percent of its staff, in July, citing a shift to an "AI Work Platform" with flatter management and smaller teams, while denying that software had directly replaced anyone who left. Meta cut roughly 8,000 jobs in May, about 10 percent of its workforce, and moved another 7,000 people into AI-focused work. On paper, Meta kept more people than it cut, which should be the model for restructuring without gutting a team. It is not working that way.
Here is the test I'd apply to tell the two apart: would an outside candidate accept the new role, at the same pay, if you wrote the job description from what the work actually needs instead of who happened to be available? If the new "AI role" is the old job with fewer hours of real content and a different name, it is a layoff wearing a reorg's clothes.
There is a real argument for the other side, and it deserves a fair hearing before I push back. If a company genuinely needs fewer people, the honest move is to say so and cut cleanly, not manufacture internal transfers nobody asked for. Monday.com's leadership priced the move specifically: a projected 19 to 20 percent revenue growth for 2026 and a margin rise from about 13 percent to about 15 percent, with most of the savings slated for reinvestment. A clean layoff with severance beats a transfer into a role that quietly dissolves in eighteen months, and dressing up a cut as a restructuring can cost more trust than the cut itself would have.
That argument is strongest when the "restructuring" turns out to be thin, and Meta's numbers show what thin looks like. The Applied AI Engineering unit, built largely from that 7,000-person transfer, reportedly reached a manager-to-contributor ratio of 50 to 1, with employees describing their work as writing coding puzzles to train models rather than engineering. Mark Zuckerberg told staff in an internal memo, reported by Reuters, that the company had "made mistakes and will almost certainly make more". Generating training data at a 50 to 1 supervision ratio is not a new job. It is the old headcount doing lower-skill work under a new label, which is exactly the gutting the restructuring was supposed to avoid.
This is not only a Meta problem. Workday's October 2026 Global Workforce Report found that 40 percent of business leaders expect AI to help their organization get more value from existing employees, against 28 percent who expect it to cut headcount outright. Most leaders say they want the redeployment path. But the same report found internal moves declining at 57 percent of employers year over year, and a separate September 2026 survey of 6,000 full-time employees found a 13-point gap between the share who know which skills they need (79 percent) and the share whose employer actually helps them build those skills (66 percent). Companies tell surveys they want redeployment and skip the training budget that would make it real.
My read, and this is opinion rather than something the data proves: most of this gap is not malice, it's cost discipline aimed at the wrong line item. Training shows up as an expense this quarter; a smaller headcount shows up as margin next quarter. A board watching quarterly numbers will underfund the thing that pays off later, even when its own executives say they believe in it.
None of this argues for a disclosure mandate, the kind Connecticut passed this year requiring employers to give written notice when automated tools meaningfully affect employment decisions. A law that forces a company to label a cut "AI-related" does not make the redeployment behind it any more real. It adds a filing. What is already enforcing this is the labor market itself: Meta's staff pushed back hard enough that Zuckerberg wrote an internal apology, and Monday.com felt the need to publicly deny that software had replaced anyone, because people were watching whether the story would hold. Competition for the employees you want to keep is a sharper check on a fake restructuring than any compliance form.
If you are restructuring a small team around AI, run three checks before calling it anything other than a layoff. Would an outside hire take the new role at the same pay, without a title rewrite built around who stayed? Did real budget and hours go into retraining, not just an announcement? Does the resulting team have a wider skill ceiling than before, not a narrower one relabeled "AI-augmented"? Pass all three and you have actually restructured. Fail any of them and you have laid people off with better branding, and your own staff will notice before your competitors do. This is the distinction we build MojoAgents around, reorganizing the work, not the people. If you are mid-reorg and want a second opinion on whether yours clears that bar, get in touch.
Sources
References used in this article. Links also appear alongside the relevant claims.
Let's make it happen.
You don't need a polished brief. A couple of lines about where you want your company to go is plenty, and we'll come back with what we'd tackle first.
We scope the work around your goals and timeline before quoting anything, so you know exactly what you're getting.