Kief Studio

CFOs Expect AI Job Cuts to Rise 9x This Year. That Work Is About to Land on Your Desk.

Build

  • Published
  • Updated
  • 8 min read
Sparse violet and gold particles drifting across black, cover for a note on AI job cuts landing work on your desk

TL;DR: In a survey of 750 CFOs, 44% planned AI job cuts in 2026, about 502,000 roles or nine times 2025's AI-attributed layoffs. Analysts say most firms do not have the AI to replace that work yet, so much of it will come back as contracts. For agencies and MSPs, the question is capacity, not demand.

AI job cuts are moving from headlines into budgets. In March 2026, a working paper built on the Duke CFO Survey put a number on what finance chiefs actually plan this year, and it was larger than last year by a wide margin. The part that matters to anyone who builds or runs technology for clients is what happens to the work after the people are gone.

Here is what the data says, what analysts think happens next, and how an agency or MSP can get ready to take on the work without a three-month hiring cycle.

Update, October 2026: we first published this note in March 2026. We have checked each figure against the survey coverage and the original publishers, removed a company example and several market statistics we could not trace to a primary source, and made the advice more specific.

What did CFOs say about AI job cuts?

Fortune reported on a National Bureau of Economic Research working paper drawn from the Duke CFO Survey, which Duke runs with the Federal Reserve Banks of Atlanta and Richmond. Out of 750 CFOs, "less than half (44%) say they plan on some AI-related job cuts." Across the economy, the co-authors estimated "just 0.4%, or about 502,000 roles out of about 125 million roles," would be lost this year, and "about half of those job losses will come from the white-collar world." (Fortune (opens in a new tab))

The nine-times comparison comes from last year's baseline. Challenger, Gray & Christmas reported that in 2025, "AI has been responsible for 54,836 announced layoff plans." (Challenger, Gray & Christmas (opens in a new tab)) Fortune rounds that to 55,000, which makes 502,000 about nine times as many.

The context matters as much as the number

John Graham, the survey's director, told Fortune: "It's not the doomsday job scenario that you might sometimes see in the headlines." (Fortune (opens in a new tab)) Half a million roles is real for the people in them and small against the whole workforce. For a firm that delivers technical work, the size of the shift matters less than where the work goes next.

Smaller firms are moving the other way

The same coverage found that "AI adoption could actually lead to an increase in hiring among smaller firms," and that "small firms noted they plan to increase hiring in technical roles." Graham's summary: "If anything, small companies are hiring a bit on the technical side." (Fortune (opens in a new tab)) So large firms trim, smaller firms build, and both need technical help they do not have in house.

Are AI job cuts actually replacing work with AI?

Often not yet. Many AI job cuts are budget decisions with an AI label. Forrester's J.P. Gownder wrote that when clients say their CEO wants to replace staff with AI, Forrester asks if they have a mature, vetted AI app ready to fill those jobs, and "nine out of 10 times, the answer is no." His conclusion: "most of the layoffs are financially driven and AI is just the scapegoat, at least today." (Forrester (opens in a new tab))

Even where the AI exists, building it is hard. RAND notes that "By some estimates, more than 80 percent of AI projects fail," which it calls "twice the already-high rate of failure in corporate information technology (IT) projects that do not involve AI." (RAND Corporation (opens in a new tab))

Put those two facts next to each other. Companies are cutting people for AI they have not built, in a field where most projects fail. The work does not disappear when the people do.

Where does the work go after AI job cuts?

Some of the work behind AI job cuts comes back. Forrester predicted for 2026: "We expect half of AI-attributed layoffs to be quietly reversed, with jobs returning offshore or at lower wages." (Forrester (opens in a new tab)) The Register's coverage of that report adds that "55 percent of employers regret laying off workers because of AI." (The Register (opens in a new tab))

Rehiring offshore or at lower wages is another way of saying the work moves outside the company. Internal roles become contracts, projects and support agreements. Some of that lands with agencies and MSPs, often as urgent requests: the integration nobody maintains anymore, the custom tool whose author left, the security review that was someone's side job.

We have covered the regret side in companies that automated the wrong thing and the rebound in why AI layoffs backfired. This note is about what a delivery team should do about it.

Why is capacity the real problem for agencies and MSPs?

Because the people who can do this work are already hard to hire. The Bureau of Labor Statistics projects that employment in computer and information technology occupations "is projected to grow faster than the average for all occupations from 2025 to 2035," with about 280,000 openings each year, and a median annual wage of $109,470 in May 2025. (Bureau of Labor Statistics (opens in a new tab))

If demand arrives in bursts after AI job cuts at your clients' clients, and every new engineer takes months to find and onboard, the agency that wins is the one that can say yes in two weeks. That is a capacity question, not a sales question.

How can an agency absorb the work without hiring?

Here is the plan I would run, in order.

1. Measure your real bench

For the next six weeks, list every billable person, their committed hours and what is left. Most agencies guess this number. Write it down. If a new 120-hour project arrived Monday, who would do it?

2. Decide what you will never staff internally

Pick the disciplines you need a few times a year but cannot keep busy full time: a security review, a data migration, a custom integration, a legacy system nobody on staff knows. Those are partner work, not hiring work.

3. Line up a build partner before you need one

Do the vetting now, not the day a client calls. Agree on how scoping works, how code and documentation get handed over, and who talks to the client. A partner who stays out of your client relationship is the point of white-label development, and we explain why that matters in your clients don't know we exist. I have also written on my own site about why many white-label arrangements fail to deliver (opens in a new tab), and Amelia, my wife and Kief Studio's co-founder and CEO, has a list of partner vetting questions founders skip (opens in a new tab).

4. Package the work you expect

Write two or three fixed-scope offers for the requests that follow AI job cuts: "take over and document this integration," "security review of what your former team built," "stabilize and host this internal tool." Clear scope turns an emergency email into a project you can start.

5. Keep security in every offer

Companies that cut IT staff still have audits, incidents and patches. When you take over a system, check access, backups and logging first. That is also the fastest way to find what the previous team left behind.

What should agencies tell clients who are planning AI job cuts?

Some of your clients will ask whether they can cut staff and let AI do the work. Answer with a short checklist rather than an opinion.

  1. Is the AI built and tested on real work? If the answer is "we plan to," the cut comes before the replacement, which is the pattern Forrester describes.
  2. Who checks the output? Every AI system needs a person who reviews what it produces. If that person is on the layoff list, the plan has a hole.
  3. What does the departing team know that is not written down? Passwords, vendor contacts, the reason a script runs at 2 a.m. Capture it before anyone leaves.
  4. What is the fallback? If the AI underperforms, who does the work next month, and at what cost?

A client who works through those four questions makes better decisions about AI job cuts, and trusts the agency that asked them. Some will slow down. Others will go ahead with a real plan, and that plan usually includes outside help for the build and the handover.

Is this only a risk, or also an opening?

Both. Agencies that solve capacity will see more inbound work from AI job cuts than they can staff internally. Agencies that cannot will spend the year turning work away. The difference is preparation done before the requests arrive, not a bigger sales effort.

It also helps to be honest with clients about AI. Many of them have heard that AI will do the job of the people they lost. Your value is building what actually works, with a person checking the output, and saying plainly what is not ready.

Frequently asked questions

How many AI job cuts do CFOs expect in 2026?

In a survey of 750 CFOs, 44% planned some AI-related cuts. The researchers estimated about 502,000 roles, roughly 0.4% of U.S. jobs.

How does that compare with 2025?

Challenger, Gray & Christmas counted 54,836 announced layoffs attributed to AI in 2025, so the 2026 estimate is about nine times larger.

Are companies actually replacing these workers with AI?

Often not yet. Forrester says that nine out of 10 times, clients planning AI layoffs do not have a mature AI app ready to do the work.

Will the jobs come back?

Forrester expects half of AI-attributed layoffs to be quietly reversed, with jobs returning offshore or at lower wages, which often means outside contracts.

What should an agency do to prepare?

Measure real bench capacity, decide which disciplines to partner for, vet a build partner in advance, and package fixed-scope offers for the work you expect.

Do smaller companies make AI job cuts too?

The survey coverage found smaller firms plan to increase hiring in technical roles as they adopt AI, while larger firms hold technical roles steady.

Where we fit

Kief Studio is a white-label build partner for agencies and studios with more client work than build capacity. Our white-label software development means we build the site, application or system and your agency delivers it under your own name. We write the code, host it when the engagement calls for it, and stay off the stage.

When the work behind AI job cuts lands as an integration, an internal tool or a system nobody documented, we can take it on without you adding headcount first. Some of our client relationships go back more than 13 years, and we would rather start with a small, well-scoped project than a big promise.

First conversation is free. No commitment.

Tell us what your bench looks like

Related articles

  1. Build

    Half of AI Layoffs Are Expected to Backfire. If Your Agency Gutted Its Tech Bench, You Can't Catch the Rebound.

    55% of employers regret AI layoffs and half may be reversed. What the research shows and how agencies can rebuild technical capacity for the rebound.

    Amelia Gagne

  2. Support

    Build the Team or Find the Partner: How to Decide Whether to Hire or Partner

    Hire or partner for technical work: what a seat really costs, when steady demand justifies a hire, what to always own, and Massachusetts contractor rules.

    Amelia Gagne

  3. Build

    How a White-Label Development Partner Works With an Agency

    How a white label development partner works with a Massachusetts agency: client communication, IP assignment, logins, hosting, standards and red flags.

    Amelia Gagne