· AI
Restructuring a Team Around AI? The Layoff-First Version Keeps Failing
More than half of employers who laid off workers for AI now regret it, and the data show why cutting first is the expensive way to restructure.
Most companies restructuring around AI are running the same experiment, and getting the same answer back. Forrester's Predictions 2026: The Future of Work report found that 55 percent of employers who laid off workers and attributed the cut to AI now regret the decision, and the firm expects half of those layoffs to be quietly rehired, according to HR Executive's coverage of the report. That is not a story about AI failing to do useful work. It is a story about treating "restructure around AI" as a synonym for "cut headcount first, sort out the org chart later," and paying twice for it.
The case for cutting first
I want to state the other side honestly, because it is not a dumb position. If a role really is redundant, carrying it while you figure out a redeployment plan costs real money every month, and severance is a known, bounded cost while a slow reorg is not. A competitor who cuts capacity the moment AI can plausibly cover it gets a cost advantage sooner, and waiting for certainty can mean waiting forever, since no vendor demo comes with a guarantee. There is a blunter version too: some "redeployment" plans are cover for a manager avoiding the harder work of managing out a performance problem. From pure cost discipline, fast can beat slow and humane.
That case holds up fine in a market where the substitute is proven. The problem is that a lot of AI restructuring is happening before it is.
What the redeploy-first data actually shows
Klarna is the cleanest public example, because it ran the experiment loudly and reported the result honestly. In 2024, Klarna eliminated roughly 700 customer service roles for an OpenAI-built assistant handling 75 percent of chats across 23 markets, and paired that with a hiring freeze that let headcount drift down 22 percent through attrition, as reported by Entrepreneur. By 2025, Klarna was recruiting human agents again, and CEO Sebastian Siemiatkowski did not blame the model. He said "investing in the quality of human support is the way of the future," conceding the AI-first approach had produced output he was willing to call lower quality once real customers started reacting to it. He blamed the decision to swap out capacity before quality held up under volume.
That pattern is not unique to Klarna. LHH's 2026 workforce survey, run across 3,000 HR leaders and more than 8,000 employees in seven countries, found 62 percent of employers now track what rehiring actually costs, and 73 percent of that group admit rehiring costs more than targeted internal mobility would have, according to LHH's summary of the findings. The same survey found 77 percent of HR leaders claim a redeployment program exists, but only 19 percent of employees recognize it. A program nobody can find is not a program, it is a slide in a board deck, and the gap is where the rehiring bill comes from.
Here is the math behind that bill. A pooled analysis of turnover research by the Center for American Progress, drawing on 11 academic studies, put the typical cost of replacing a worker at about 21 percent of annual salary for most roles, and up to 213 percent for highly skilled positions, once you count the vacancy, the search, and the six to twelve months a new hire needs to reach full productivity, per CAP's analysis. Cut a $60,000 role, save $60,000 for a year, then discover the AI system covers most of the work but not the edge cases customers actually call about. Rehiring costs roughly $12,600 in pure replacement overhead, on CAP's 21 percent median, before you pay a cent of the new salary, on top of whatever revenue or trust you lost in the gap. That is a fifth of a year's savings spent just to get back to where you started, and it assumes you can find the same person, which Klarna could not: it recruited a fresh pool of students and remote workers instead.
The actual case for Mojo's approach
None of this means AI cannot genuinely shrink a team's workload. It means the order of operations matters more than the announcement. Measure what a workflow costs, run the AI system against real volume long enough to see where it breaks, and only then decide whether the freed-up hours mean fewer people or the same people doing higher-value work already backlogged. Almost every small business I have worked with has real work nobody has time for, which makes redeployment an easier call than most executives assume: you are not inventing work to justify a headcount, you are finally getting to work already worth doing.
This is the part of MojoAI that has nothing to do with picking a model. Restructuring around AI is an organizational design problem with a technology component, not the other way around, and the companies getting burned right now let the technology component make the org chart decision by itself. If your reorg is being driven by a vendor's roadmap more than your own numbers, let's work out what the workflow actually costs before you cut anything.
Sources
References used in this article. Links also appear alongside the relevant claims.
Tell us what's on your mind.
You don't need a polished brief to reach out. A two-line email about what's bugging you is plenty; we'll tell you straight if we're the right fit, and what we'd tackle first.
We'll scope the work around your workflow, goals, and timeline before quoting anything, so you know what's included before committing.