The AI headcount saving comes with a buyback clause

Gartner says three in ten people cut because AI took their job will be rehired by 2029, at a higher price. In India one in four CHROs is already planning the cut. Nobody is pricing the rehire.

Gartner has put a number on something most founders already suspect. By 2029, 30 percent of employees laid off because AI replaced them will have to be rehired, often at significantly higher cost^[1]. The same note adds that by 2027, 75 percent of organisations that book AI productivity gains as cost savings will be overtaken by rivals that put those gains back into upskilling and modernisation^[1]. Read the two lines together and they describe a loan, not a saving.

The Indian numbers say the loan is being taken out now. A Great Place To Work India study finds nearly 25 percent of CHROs expect AI to reduce their workforce by 1 to 20 percent within two years, while 45 percent say it is too early to tell^[2]. The same study puts 13 percent of employees in the high-burnout band^[2]. So a quarter of HR heads are planning a reduction, half do not know, and the people who stay are already carrying more than they can hold.

How it plays out in a 200-person company

An agent handles most of the routine tickets in operations, so eight people go. The saving looks clean on the monthly P&L. Six months later the exceptions start piling up, because the agent handles the routine and the routine was never the expensive part. The person who knew why a particular client gets invoiced differently is gone. You go back to the market for someone with the same context, and the market has moved. Gartner’s phrase for this is significantly higher cost^[1]. In practice it is a notice period, a hiring fee, 90 days of ramp and a CTC that no longer sits inside the band the departed person was on.

AI removes tasks, not jobs

The reason the rehire happens is that AI removes tasks, not jobs. What is left of the job after the tasks go is judgement and exception handling, and that is exactly what the mid-level people who got cut were doing. The market knows this even when the cost case does not. Among Indian employers adding staff this quarter, 62 percent cite changing roles and skills as the main reason for hiring^[3]. The same companies that deleted a role in March are hiring for it in October under a new title.

The fresher end makes it worse. Deloitte India finds 85 percent of employers now use AI somewhere in hiring but only 35 percent say it has created new entry-level roles or skill needs^[4]. So the pipeline that would have grown into the mid-level roles is being throttled at the same time the mid-level roles are being cut. Do both for two years and in 2028 there is nobody in the building who knows how the process actually works.

Do the arithmetic properly

None of this is an argument against using AI in operations. It is an argument for doing the arithmetic properly. Before any AI-driven reduction, write down who holds which process knowledge and who owns the exceptions after they leave. Put Gartner’s 30 percent into the business case as a haircut on the saving, and price the rehire at next year’s market, not this year’s band. Then take a share of what is left and spend it on the people who stay, because Gartner’s second prediction is that the companies reinvesting the gain are the ones still standing in 2027^[1].