Only 11% of 775 Leaders Put AI in Workforce Plans

A Conference Board survey of 775 leaders finds only 11% have built AI into enterprise-wide workforce and financial plans. Here is how to build a 2027 headcount plan that starts from redesigned workflows.

By Rajesh Beri·November 20, 2026·8 min read
Share:
A conference room table with a printed 2027 headcount spreadsheet, a laptop open to an org chart, red pen markings crossing out and re-adding rows, and two coffee cups from a CFO and CHRO planning session.

Illustration generated using AI

If your 2027 plan books AI savings as fewer hires, you are probably ahead of the evidence, and most of your peers have not written the number down at all. A Conference Board survey of 775 leaders, released November 18, found that only 11% have built AI's expected effects into workforce and financial plans across the enterprise. Another 37% say they will buy AI capability instead of adding headcount. A plan that turns that intent into a defensible number starts from redesigned workflows and a record of where saved time actually goes, because removing a task often creates work somewhere else.

The researchers say so themselves. "Eliminating one task can create new responsibilities elsewhere," Matt Rosenbaum, principal researcher in the Conference Board's US human capital practice, said in the release. The full report, AI & Workforce Planning: How Leaders Can Navigate Tech-Driven Change, is gated to members, so the figures below come from the public release and the trade coverage of it.

What Share of Companies Have Put AI Into Their Workforce Plans?

About one in nine. The release splits the 775 respondents into four groups:

Stage Share
AI effects built into workforce and financial plans enterprise-wide 11%
Built in for some functions, roles, business units or geographies 35%
Discussing the workforce implications, nothing in the plans yet 34%
Not yet assessing the implications 13%

The study combined the survey with 35 interviews. The release gives no field dates and no regional split beyond "international leaders", so read the split as directional. The board's own Q3 CHRO Confidence Index, a separate survey of US CHROs, put the enterprise-wide figure at 19% in October; the 11% comes from a broader international sample.

The middle two rows are where most readers sit. A third of companies have AI in some plans, often in one function or business unit, and another third are talking about it without a number in the spreadsheet. Canadian HR Reporter's write-up rounds the headline to "just 1 in 10 employers". Either way, the CFO's 2027 model and the CHRO's 2027 headcount plan are mostly being built on separate assumptions.

What Leaders Expect AI to Do to Headcount

Most expect a modest cut, and almost none expect growth. Over the next two years, per the release, 39% expect AI to slightly reduce full-time employment, 8% expect significant reductions, 32% expect little or no net effect, 5% expect an increase and 17% say it is too early to estimate. Taken together, 47% expect fewer full-time employees.

The actions they plan point somewhere less dramatic. Nearly 70% expect redesigning jobs, tasks or workflows to be among their most impactful AI workforce moves, and 68% call upskilling or reskilling a top priority. Only 17% name cutting full-time headcount in AI-affected areas as a top action. The 37% who plan to invest in AI instead of adding people describe a hiring freeze more than a layoff.

That gap matters for your plan. A forecast that assumes reductions while the operating actions are redesign and reskilling will book savings before the work that produces them is done.


Why Saved Time Does Not Turn Into Fewer Hires

Time freed by AI rarely leaves the building on its own. The Conference Board's point is that a task an agent removes tends to push review, exception handling or coordination onto someone else. "Companies can no longer rely on traditional headcount forecasts when the work itself is changing so rapidly," Diana Scott, who leads the board's US Human Capital Center, said in the release.

The pattern shows up in the cases this site has tracked. Engineering teams using coding agents shipped more pull requests and then hit a code review bottleneck. Meta pushed spans to 50 people per manager and then wanted managers back for the work AI did not absorb. Apple paused 5,000 AppleCare layoffs tied to its support agent. Genpact's survey found 92% kept or grew headcount while adding agent fleets, with new roles to run them.

The cost of getting it wrong is measured. In an April 2025 Orgvue survey of more than 1,000 leaders, about 4 in 10 had laid people off because of AI, and 55% of those said they made the wrong decisions. A February 2026 survey of 600 HR leaders who had overseen layoffs, run by the outplacement firm Careerminds and reported by IBTimes UK, found 30.9% spent more on rehiring than automation saved, and another 42.4% roughly broke even. Forrester's 2026 predictions expect half of AI-attributed layoffs to be quietly reversed, with the jobs returning offshore or at lower wages.

The strongest case for cutting early is real: if you wait for perfect evidence, a competitor with a leaner cost base prices you out. Forrester's own jobs forecast still projects AI could account for 6% of total US job losses through 2030, about 10.4 million roles. The same release warns that many firms announcing AI layoffs "do not have mature, vetted AI applications ready to fill those roles." If the agent that is supposed to absorb a role is still in pilot, a cut to that role belongs in 2028's plan.

How to Build a 2027 Headcount Plan Around Redesigned Work

Start from the workflow, then decide what the time is worth, and only then decide headcount. The Conference Board's six recommendations run in that order: look beyond headcount, prioritize work redesign, turn AI productivity into business value, build a more flexible workforce, make workforce planning continuous and strengthen cross-functional collaboration.

The report names three tools for this, a Workflow Realization Screen, a Work Transition Ledger and a Value Capture Bridge. The details sit behind the member paywall, but the release describes them as steps: find where AI can be deployed, track how it changes work across a workflow, then test whether those changes produce a measurable business benefit. You can run a version of each without the report.

A work transition ledger is a per-workflow record of every task AI changed, the hours freed, and where those hours went: absorbed into new review work, redeployed to other output, removed from the hiring plan, or unaccounted for. Watch the last category. Hours with no destination should stay out of the savings line until someone can say where they went.

The value test is a single question per workflow: does the freed time support higher output, better service, reduced hiring needs or redeployment? Each answer goes into the plan differently. Higher output belongs in the revenue line. Reduced hiring needs belong in the hiring plan. Redeployment needs a reskilling budget, which is where the 68% priority lands. If finance already models scenarios in a planning tool such as Pigment, build the ledger categories in as drivers so the headcount line moves only when a workflow's ledger does.

The ownership question is the one most plans skip. When AI shows up in some plans but not others, it is usually because one function owns the pilot and nobody owns the number. Scott's prescription in the release is that HR, business, finance and technology leaders work out together where AI creates value and where to invest in talent. In practice that means one named owner per workflow line, signed off by both the CFO and the CHRO.


What to Do Before Your 2027 Plan Locks

This Week:

  1. Ask finance and HR separately what AI savings each has assumed for 2027. If the numbers differ, or one of them is blank, you are in the 35% or the 34%.
  2. Pick the three workflows where AI is already in production and list who reviews its output today. That review work is the first entry in your ledger.

This Month:

  1. Build a work transition ledger for those three workflows: tasks changed, hours freed, and where each hour went. Leave the "unaccounted" column visible.
  2. Convert any planned AI headcount reduction into a hiring freeze on the affected roles instead of a cut, until the ledger shows the hours leaving. That matches what the 37% plan to do, and it is reversible.
  3. Name one owner per workflow line in the plan, with CFO and CHRO sign-off.

Before the Plan Locks:

  1. Price the reskilling. If 68% of your peers rank it a top priority, your redeployment line needs a training budget attached to it, by role.
  2. Set a quarterly review of the AI assumptions in place of the annual forecast. The Conference Board's recommendation is continuous planning, and agent capability changes faster than a budget cycle.

The Bottom Line

The first wave of AI headcount decisions has already been graded. Orgvue's respondents who cut staff for AI mostly say they got it wrong, and in the Careerminds survey, by IBTimes' account, only about a quarter came out ahead after rehiring costs. The Conference Board's data says most companies have not yet put a 2027 AI number in writing, which leaves room to do it in the right order. Fill in the ledger for each workflow before you enter a savings figure, and give that figure a named owner.

Continue Reading

Share:

Frequently Asked Questions

What share of companies have built AI into their workforce plans?

In The Conference Board's November 2026 survey of 775 leaders, 11% had incorporated AI's expected effects into workforce and financial plans across the enterprise. 35% had done so for some functions, roles, units or geographies, 34% were only discussing it and 13% were not yet assessing it.

How many leaders expect AI to reduce headcount in the next two years?

Per the Conference Board release, 39% expect a slight reduction in full-time employment and 8% expect significant reductions. 32% expect little or no net effect, 5% expect an increase and 17% say it is too early to estimate.

Why doesn't time saved by AI translate directly into fewer employees?

Removing a task often creates review, exception handling or coordination work elsewhere. The Conference Board's Matt Rosenbaum put it as: eliminating one task can create new responsibilities elsewhere. A headcount cut only holds if the freed hours actually leave the workflow.

What is a work transition ledger?

It is a per-workflow record of each task AI changed, the hours freed and where those hours went: new review work, redeployment, fewer open requisitions, or unaccounted for. The Conference Board report names a Work Transition Ledger as one of its three planning tools.

Do companies regret AI-driven layoffs?

In an Orgvue survey of more than 1,000 leaders, about 4 in 10 had laid people off because of AI and 55% of those said they made the wrong decisions. A Careerminds survey of 600 HR leaders found 30.9% spent more on rehiring than automation saved.

Newsletter

More on AI ROI & Cost, every Sunday

The week's enterprise AI news, AI ROI & Cost included, in one Sunday email.

Weekly, free. Unsubscribe any time. Privacy

Latest Articles

View All →