AI Readiness for CEOs: What the Mercer 99-to-32 Gap Reveals

Most CEOs already believe AI will reshape their workforce. Almost none believe their organization can actually pull it off. That gap is the real story of AI readiness for CEOs in 2026, and it is wider than any board deck admits.

Mercer’s Global Talent Trends 2026 put a number on it. 99 percent of CEOs expect AI and automation to drive headcount reductions within two years. Only 32 percent believe their organization can integrate human and machine capabilities well (Mercer, 2026). Near-total conviction about the destination. Near-total doubt about the ability to get there.

I keep seeing the same conversation in the boardroom. The CEO is convinced AI will change the cost structure. The same CEO cannot tell you whether the company is built to absorb that change. Conviction is everywhere. Capability is rare.

The danger is not the ambition. Ambition is fine. The danger is acting on ambition you have never pressure-tested. A reduction in force is the most irreversible decision a CEO makes, and 99 percent are planning one on the strength of a belief about a capability only 32 percent claim to have.

The 99-to-32 Gap Is a Capability Problem

Read the two Mercer numbers together and the meaning is hard to miss. The plan is set. The ability to execute it is not. A workforce reduction without the capability to redeploy the people who remain is not a strategy. It is a bet that automation will cover a gap you have not measured. Ambition is not a plan. A plan names the capability it depends on, then confirms that capability actually exists before the budget moves.

This pattern shows up in every serious readiness study. Cisco’s AI Readiness Index 2025 found that only 13 percent of organizations are fully ready to deploy AI, and that figure has stayed flat for three years across 8,000 business leaders (Cisco, 2025). Ambition climbs every year. Readiness does not move.

If you cannot say where your organization sits on that capability curve, fix that before any headcount decision. A free 60-second assessment at Elevates.AI/launchpad gives you a readiness baseline across data, governance, and workforce, which is the same capability gap the Mercer number describes.

Why AI Readiness for CEOs Starts With Capability, Not Conviction

Most executive AI conversations start at the wrong end. They start with what AI could do. AI readiness for CEOs starts with what your organization can currently absorb. Those are different questions, and only the second one predicts whether the investment returns anything.

McKinsey’s State of AI found that 88 percent of organizations use AI in at least one function, but only 39 percent report measurable enterprise impact (McKinsey, 2025). The adoption barrier collapsed. The value barrier did not. A CEO who treats access to AI as readiness is measuring the wrong thing.

Capability is specific. It is whether your data is governed, whether your teams are trained, whether you have an escalation path when an AI decision goes wrong, and whether someone owns the outcome. Conviction does not survive contact with any of those questions.

The CFO sees this gap before anyone else, because it shows up in the numbers. An AI strategy for CFO leaders is a question about return, not enthusiasm. When 88 percent adopt and 39 percent see impact, the spread is wasted spend sitting on a balance sheet. Readiness is the variable that decides which side of that spread you land on, which is why the finance function should treat a readiness baseline as due diligence, not a nice-to-have.

The Confidence Is Already Falling

Here is the part that should worry any board. Executive confidence is not climbing toward the ambition. It is dropping. Mercer found that 51 percent of C-suite leaders feel well prepared to succeed in the human-machine era in 2026, down from 65 percent in 2024 (Mercer, 2026). The people closest to the strategy are growing less sure, not more.

The workforce feels it too. Employee concern about losing a job to AI rose to 40 percent in 2026 from 28 percent in 2024, and only 44 percent of employees say they are thriving at work, down sharply from 66 percent in 2024 (Mercer, 2026). A frightened workforce does not adopt new tools well. The readiness gap and the trust gap are the same gap.

There is a feedback loop hiding in these numbers. Lower confidence leads to hesitation, hesitation leads to half-funded initiatives, and half-funded initiatives produce the weak results that lower confidence further. Breaking the loop requires evidence, not encouragement. A measured readiness baseline gives a leadership team something concrete to act on, which is the only thing that turns falling confidence around.

The order matters here. You measure capability, then you decide on workforce changes, not the reverse. A readiness assessment that maps your gaps before you cut is the difference between a plan and a guess. You can run that baseline in about a minute at Elevates.AI/launchpad and see exactly which capabilities are ready and which are not.

What CEOs Should Measure Before They Cut

If 99 percent of CEOs are planning reductions, the responsible move is to measure the capability to execute them first. Four questions separate readiness from ambition.

Can your data support the automation you are planning, or is it governed in name only. Do the people who remain have the skills to work alongside the systems you are deploying. Is there a governance path for when an automated decision is wrong, because there will be one. And does a named executive own the outcome, not just the announcement.

Notice what these questions have in common. None of them are about the AI itself. They are about the organization around it. That is the part executives consistently underweight, because the technology is the exciting part and the plumbing is not. The plumbing is what determines whether the technology returns anything. A model is only as ready as the data feeding it, the people running it, and the controls catching it when it fails.

Most organizations cannot answer all four with evidence. That is the starting point, not a failure. The companies that close the 99-to-32 gap turn those four questions into a measured baseline and a sequenced plan. Our breakdown of how the major AI maturity models compare shows where most executive frameworks stop short of this.

What the 32 Percent Do Differently

The 32 percent that can integrate human and machine capabilities are not better funded. They are better sequenced. They measured readiness before they committed to a number, so their automation plans rest on capability they have confirmed, not capability they assume.

Three habits show up consistently. They treat data governance as a precondition, not a cleanup task that happens after deployment. They map workforce skills against the roles AI will change, so redeployment is a plan instead of a layoff. And they assign a single accountable owner for AI outcomes, which turns a board ambition into a tracked program with milestones.

None of this is exotic. It is the difference between deciding and then checking, versus checking and then deciding. The 99 percent set the ambition first and discover the gaps in production. The 32 percent find the gaps first and close them before they scale. Same companies, opposite order.

The Human-Machine Equation Runs Through the Workforce

Mercer’s framing is deliberate. The gap is about integrating human and machine capabilities, not deploying machines alone. That means the readiness question is as much about your people as your platforms. You cannot redeploy a workforce you cannot see.

There is a hard version of this that most plans ignore. Automation does not remove the need for judgment. It moves it. The people who remain are asked to supervise systems, catch errors, and handle the cases the model cannot. That is a different skill set than the one most teams have today, and it does not appear on its own. Without a deliberate plan to build it, the 32 percent capability stays out of reach no matter how much you spend on tools.

Knowing the skills you have, the skills you lack, and the roles most exposed to automation is the input the 32 percent have that the others do not. Levos.AI, our sister platform, focuses on exactly that workforce intelligence layer. Disclosure: Levos.AI is owned by the same team behind Elevates.AI. The point is not the tool. A headcount decision made without a skills map is the 99 percent talking. A decision made with one is the 32 percent.

Where AI Readiness for CEOs Actually Starts

The Mercer number is a warning, not a verdict. The 99 percent who expect AI to cut headcount and the 32 percent who can actually integrate human and machine work are mostly the same companies, separated only by whether they measured capability before they acted. If you are planning AI driven change and cannot yet prove your organization can absorb it, start with the gap, not the cut. Run the free assessment at Elevates.AI/launchpad, get your readiness baseline, and decide from evidence instead of conviction.

Frequently Asked Questions