The Boomerang Workforce: Gartner Says 30% of AI-Replaced Workers Will Be Rehired. Is That Too Low?

Boomerangs come back. That is the entire point of a boomerang.

For about two years, the most reliable way for a CEO to please a room full of analysts was to say “AI” and “headcount” in the same sentence, preferably with a downward arrow. The invoice for that sentence is now arriving.

On October 28, at its HR Symposium in Orlando, Gartner put a number on the hangover: by 2029, 30% of employees who were terminated and replaced by AI will be rehired, often at a higher cost. The part of the headline worth reading twice is the conditional tacked on the end: if organizations don’t overhaul workforce planning.

What Gartner actually said

The October 28 release is, at heart, a launch vehicle for a new Enterprise Strategic Workforce Planning framework presented by Caitlyn McDonough, VP Analyst. The supporting data:

• A July 2026 Gartner survey of 297 CHROs found only 19% of organizations have workforce planning that actually helps leaders understand how work is changing, accounts for AI, and realigns talent with business goals.
• Organizations that do plan well are 1.4x more likely to beat revenue goals and 1.5x more likely to beat cost goals.
• The framework runs top-down: identify the capabilities critical to strategy, forecast future work demand, assess future supply (human and AI capacity), quantify the gaps and weigh fixes by cost, risk and speed, then govern how it all gets funded. McDonough’s summary line: leaders should “stop planning for the workforce they have” and plan for the one their strategy requires.

The 30% figure itself is not new. It debuted on September 9 in Gartner’s “four shifts shaping the future of work” release, fronted by VP Analyst Tori Paulman, and the follow-up interviews are more revealing than either press release:

• Paulman told Computerworld the 30% is an average across company sizes; for large enterprises she estimates the boomerang closer to 40%.
• Gartner analyzed more than 1 million layoffs in 2025 and found cuts genuinely driven by AI productivity (one worker made so productive that fewer were needed) were under 1% of the total.
• Of the AI-attributed layoffs in the first half of 2025, 17% were really commercial pivots, which Paulman flagged as possible AI-washing.
• More than half of Gartner’s enterprise clients have been handed a savings target to hit through AI. Paulman described executives as steeped in “magical thinking.”
• A companion prediction: by 2027, 75% of organizations that bank AI productivity gains as cost savings will be eclipsed by competitors that reinvest them.

In other words: the same research firm telling you the cuts were premature also found that almost none of them were caused by AI doing the work. Hold that thought; it matters later.

The scoreboard: what other public sources say

Gartner is not out on a limb here. If anything, it is one of the more conservative voices in the room.

And then there is the poster child. Klarna said in early 2024 that its OpenAI-built assistant was doing the work of 700 customer service agents. By 2025 CEO Sebastian Siemiatkowski conceded the push had produced “lower quality” support, and the company began recruiting humans again through an Uber-style remote pilot. One detail rarely makes the headline: Klarna shrank largely through a hiring freeze and attrition, not mass layoffs. The most famous “AI replaced 700 people” story was mostly a story about not backfilling. Ford, IBM, Booz Allen Hamilton, Alphabet and CSX have all walked back cuts or announced rehiring drives as well.

The wave the boomerang is coming back from

Four months of 2026 out-cut all of 2025, and AI has been the top stated reason for layoffs since March. That is the pool Gartner’s 30% will eventually be measured against. Keep the word stated in mind: Challenger counts what companies say in their announcements, and the firm itself has noted how fuzzy the “AI” label is in practice.

So, do I agree with Gartner?

Mostly, and for slightly different reasons than Gartner gives.

The core claim, that a large share of AI-justified cuts were premature and will be partly unwound at a premium, is about as well-supported as any forecast in this space gets. Every independent data point points the same way: regret rates north of 50%, rehiring already underway at roughly a third of companies, and a majority of firms that rehired either losing money on the round trip or merely breaking even. Robert Half’s numbers suggest the 30% mark is arguably being approached now, in 2026, three years ahead of Gartner’s deadline.

Where I part ways is the number itself, and what it implies about the mechanism. “30% rehired” reads as if AI tried, failed, and the humans came home. The evidence describes something messier: a lot of budget cutting wearing an AI costume, followed by a quiet refill of the work through cheaper channels. That distinction matters, because the fix for “AI isn’t ready” is different from the fix for “we used AI as an excuse.”

Five problems with the 30% number

1. The denominator is contaminated

Gartner’s own analysis found that genuine AI-productivity cuts were under 1% of 2025 layoffs. Forrester calls the rest of the pattern “AI washing”: financially driven layoffs announced as AI-driven, at companies without mature AI ready to do the work. Yale Budget Lab’s Martha Gimbel has made the same point more bluntly, noting that blaming AI is easier for a CEO than admitting trouble with tariffs, immigration or the post-pandemic hiring hangover. Andy Challenger put it neatly: whether or not AI is replacing the jobs, it is replacing the budget for them.

If most “AI layoffs” were really cost cuts, then rehiring is not evidence that AI failed. It is evidence that the work never went away. That is a workforce planning failure, which, to be fair, is exactly Gartner’s argument.

2. There is no baseline

Companies rehire after ordinary layoffs too. Frank Dickson of Dickson Research pointed out that without a comparison rate for non-AI layoffs, the 30% is “a figure floating with no anchor.” Info-Tech’s Valence Howden went further and questioned whether the number can be substantiated at all, while still agreeing the cuts were excessive. A 30% reversal rate is damning only if normal restructurings reverse at, say, 10%. Nobody has published that comparison.

3. “Rehired” is not the same as “reversed”

Forrester expects reversed roles to come back offshore or at significantly lower salaries. Dickson lists the other disguises: contractors, retitled positions, and workloads silently absorbed by whoever survived. None of those show up as a rehire, so the true reversal rate is likely higher than 30%, while the share of laid-off people who actually get their jobs back is likely lower.

Gartner says rehiring will cost more; Forrester says the jobs will come back cheaper. Both can be true at once: cheaper per head, more expensive in total once you add recruiting, ramp-up time, rework and the knowledge that walked out the door. SHL’s finding that half of AI-replaced workers want a substantial raise to return is the per-head half of that equation for anyone you do want back.

4. Many of the surveys come from people who sell the cure

Careerminds is an outplacement firm. Orgvue sells workforce-planning software. Gartner launched this prediction (twice) alongside a workforce-planning framework at an HR conference. None of that makes the data wrong, and the results line up across very different methods. But it is worth noticing that nearly every alarm about AI-layoff regret is being rung by someone with a product for the regret.

5. The conditional makes it hard to grade

“30% will be rehired if organizations don’t overhaul workforce planning” is a prediction with a built-in alibi. If 2029 arrives at 20%, clients planned better. If it arrives at 45%, they didn’t plan enough. Meanwhile, Forrester’s harder-edged forecast (half reversed by the end of 2026) will be checkable within weeks. Aikido Security’s CISO Mike Wilkes summed up the stakes well in CIO’s coverage: even if 30% is right, it is an enormous error rate for a strategic decision, the kind no CFO would tolerate on any other capital expense.

The angle most coverage misses: institutional knowledge is a security control

Anyone who has worked in security operations will recognize the part of this story the HR framing undersells. Every organization runs on controls that are not written down: the analyst who knows why a firewall exception exists, the engineer who remembers which legacy service still authenticates with a shared secret, the support agent who can tell a social-engineering call from a confused customer in about four seconds. That knowledge is not documented because the person is the documentation.

Cut that person and you have not reduced headcount. You have silently removed a control, and you will not find out until the incident review. Careerminds found that a third of HR leaders lost critical skills after AI-driven layoffs, and that almost nobody waited more than a year to start rehiring. In security terms, that is a detection gap closing itself the expensive way.

The irony is sharpest with agentic AI. The more autonomy you hand to software, the more you need people who understand the systems well enough to supervise it, review its output and recognize when it is confidently wrong. One analysis CNBC cited noted that companies pushing automation often cut the very people needed to oversee the AI. Howden at Info-Tech expects rehiring to accelerate as AI governance requirements grow. Paulman’s own prescription is to use the technology in a verifiable, correct way, which is a sentence that quietly requires a human to do the verifying.

What leaders should actually do

Gartner’s framework is more sensible than its headline. Strip away the conference packaging and the advice is sound:

1. Prove the capacity before you cut the people. An AI system in production, measured on real work, is a supply line in a workforce plan. A vendor roadmap is not. Pilot, measure, then reduce; never the reverse.
2. Audit your own “AI layoffs” honestly. If a cut was really budget-driven, call it that internally, so the planning that follows is aimed at the right problem.
3. Plan around work, not job titles. Gartner’s top-down approach (strategy, then critical work, then human plus AI supply, then gaps) is the right order. Bottom-up “which tasks could AI do” exercises are how organizations talk themselves into cuts the work does not support.
4. Capture tacit knowledge before anyone leaves. Runbooks, decision logs, recorded walkthroughs. Treat undocumented expertise as a risk register item, not a nice-to-have.
5. Price the boomerang up front. Put rehiring premiums, ramp time, offshore transition costs and quality loss into the business case. Careerminds’ data suggests most companies that skipped this step broke even at best.
6. Reinvest some of the gains. Gartner’s prediction that cost-savers will be eclipsed by reinvestors by 2027 is the most interesting number in the whole package, and the least quoted.

The boomerang was never a failure of the technology. It was a failure to read the throwing instructions. AI will keep reshaping work, probably more than it has so far. But organizations that treat it as a reason to plan carefully will spend the next three years compounding gains, and the rest will spend them on recruiters.

Sources

• Gartner: 30% of employees replaced by AI will be rehired by 2029 if organizations don’t overhaul workforce planning (Oct 28, 2026)
• Gartner HR Symposium/Xpo Orlando, Day 3 highlights
• Gartner: Four shifts shaping the future of work (Sept 9, 2026)
• CIO Dive: Gartner on AI-replaced workers being rehired
• CIO: Layoff remorse, Gartner says one in three AI-eliminated positions will be restored
• HRD: AI rebounding will cost more (Careerminds, SHL data)
• CPA Practice Advisor: 9 in 10 HR leaders regret AI-driven layoffs (Careerminds)
• The Register: Forrester predicts half of AI layoffs quietly rehired at lower pay
• Forrester Predictions 2026 release (PDF)
• People Matters: 55% of employers regret AI layoffs (Forrester, Orgvue, Careerminds, Robert Half)
• Challenger, Gray & Christmas: May 2026 job cuts report (PDF)
• Challenger, Gray & Christmas: July 2026 job cuts report (PDF)
• CBS News: AI emerges as a top cause of layoffs
• Fortune: Yale Budget Lab report raises questions of AI-washing
• NewsNation: Yale/Brookings and Stanford findings on AI and jobs
• Ohio CPA: Chicago Fed finds AI-exposed jobs grew 4%
• Computing: MIT report finds 95% of GenAI pilots fail to deliver returns
• Entrepreneur: Klarna hiring human agents again
• The Daily Star: Klarna’s AI-first shift and attrition