Gartner: 30% of AI Layoffs Will Be Rehired by 2029. A Checklist Before You Cut

Gartner predicts that by 2029, 30% of employees laid off due to replacement by AI will need to be rehired, often at a significantly higher cost. For large enterprises the figure is closer to 40%. The problem is rarely the technology. It is what the company loses when the person leaves: knowledge that was never written down.

What Gartner predicted

Gartner published the prediction on 9 September 2026 as part of four shifts it sees shaping the future of work.

• The prediction: by 2029, 30% of employees laid off due to AI replacement will need to be rehired, often at a significantly higher cost.
• Enterprises: Gartner VP analyst Tori Paulman estimated roughly 40% for large organizations.
• The reasons: layoffs "deplete talent pipelines and erode institutional knowledge," cuts can go too deep and too soon, and with labor force growth flat or declining, recruitment, training and onboarding costs rise.

Paulman's summary: "When business and IT executives look back on the early AI era, they will realize their greatest mistake was believing that work automation was the point, when workforce amplification was the opportunity."

Why the rehire bill is easy to miss

The saving and the cost show up in different places. The saving is announced in the year of the cut, often on an earnings call. The rehire happens later, on a recruiting budget, sometimes under a different manager. One is visible. The other rarely makes the news.

For a large company, a rehire is expensive. For a ten-person business, it may not be possible at all. If the one person who knows how your best customers like to be handled leaves, that knowledge leaves with them.

What should you check before cutting a role for AI?

Use this checklist before any AI-driven change to a role. If you cannot complete it, you are not ready to make the decision.

1. Customer knowledge: which customers need special handling, and what does "special" mean for each?
2. Exceptions: what does this person do when the normal process does not fit? List the last ten times it happened.
3. Relationships: which suppliers, partners or contacts do they call directly, and why them?
4. Workarounds: which tools or processes do they quietly patch every week?
5. History: which past decisions, disputes or promises only they remember?
6. Judgment calls: which decisions do they make that are not written in any policy?

Automate the task first, decide about the role later

Step	What to do	How long
1. Document	Complete the checklist above with the person, not about them	1 to 2 weeks
2. Automate a task	Hand one routine task to AI while the person stays in the loop	Ongoing
3. Measure	Track errors, handoffs and customer complaints	One quarter
4. Decide	Change the role based on data, not a forecast	After step 3

The better use of AI in a small team

Using AI to do the work is not the mistake. Cutting the person before you know which parts of the work AI can actually hold is. We saw a version of this at scale this week, when Apple kept its AI support assistant but paused a plan to cut 5,000 AppleCare staff.

The better route for most small businesses is the one Gartner calls amplification: automate the repetitive tasks so the people who hold your knowledge can spend more time using it.