Adoption is near universal. Financial return is not. The gap between the two closes in one place: the CEO's office, and the leadership team built around the work.
Almost every large company now runs AI somewhere in the business. Almost none of them show the result in earnings. Ten years of McKinsey survey data, three years of agentic pilots, and billions in enterprise spend have produced a single durable finding: the constraint is no longer the model. The constraint is leadership design.
Christian & Timbers places the executives who close this gap. What follows is our read of the evidence on how CEOs are leading through the shift, and what the evidence implies for the people a board needs in the room.
The value gap, in numbers
McKinsey's tenth annual State of AI survey, fielded May 4 to June 8, 2026 across 1,719 leaders in 97 countries, puts hard figures on the problem:
- 89% of organizations use AI regularly in at least one business function
- 80% of users report gains in their own productivity
- 37% report any positive contribution to EBIT, flat against 2025
- Roughly 6% qualify as high performers, with more than 5% of EBIT attributable to AI
- 40% of enterprises above $1 billion in revenue are scaling agents, up from 27%
- 32% declined to buy at least one software product because agentic coding tools let them build in-house
- 39% expect AI-related headcount declines in the coming year, against 14% who saw declines in the past year
Four in five employees feel faster. Barely one in three finance functions sees the difference. Everything of consequence for a CEO sits between those two numbers.
80% business, 20% technology
Eric Kutcher, McKinsey's North America chair, describes the work as 80% business transformation and 20% tech transformation.
The ratio explains the flat EBIT line. Most enterprises have run the 20%. Tools were licensed, platforms were selected, pilots were funded, and good things happened. The good things did not add up, because the surrounding process stayed intact.
Three findings from McKinsey's survey work point the same direction:
- Workflow redesign shows the strongest correlation with EBIT impact of any organizational attribute tested.
- CEO oversight of AI governance shows the strongest correlation at companies above $500 million in revenue. Only 28% of organizations report a CEO in the role. Only 17% report board-level ownership.
- CEO-led digital transformations are roughly 1.5 times more likely to succeed than programs led primarily by technology teams.
Delegating the AI agenda to a CTO produces a technology program. AI value comes from a business program with a technology component.
Five moves the leading CEOs are making
1. They build, instead of being briefed
The old model of technology leadership was sponsorship: fund the investment, ask sharp questions, trust the experts. Agentic systems break the model, because the unit of work stops being a tool and becomes a living workflow. McKinsey's review of more than two dozen AI transformations found adoption outpacing leadership readiness, with the fastest organizations pushing executives to build agents themselves. One immersion format has leaders build an agent for a personal problem, purely to produce the moment when AI stops being a concept and starts being a collaborator.
A leader who has hit the limitations firsthand sets expectations without overselling or catastrophizing. Credibility of this kind resists faking.
2. They set an ambition large enough to break the old process
Incremental targets get met with the existing process plus a copilot. The CEOs producing enterprise-level results are naming outcomes the current process cannot deliver: triple the share price in four years, triple the customer base in one function, cut cycle time by an order of magnitude. The ambition forces the redesign. A modest target never does.
3. They redesign the workflow before deploying the agent
End-to-end redesign covers roles, decision rights, handoffs, governance, and sequencing. Companies applying AI to an unchanged process inherit the constraints of the process. Companies rebuilding the process around human and agent capability inherit a different cost curve.
4. They rebuild the management layer for human and agent teams
Managers increasingly orchestrate work across people, agents, and systems. Output arrives faster and with wider variance in quality, which shifts the managerial job toward discernment: is the answer correct, is the answer relevant, will the team defend the logic, where does the risk sit. Teams will bring leaders low-quality AI output, and managers need the judgment to catch the problem and coach the person who produced the work.
Practical mechanisms exist. One global life sciences group redesigned the weekly operating review with a rotating challenger role, briefed to argue against the leading conclusion and to log what AI proposed against what humans decided. Speed increased. Accountability survived.
5. They own governance and the life cycle of the agent estate
An enterprise running 15,000 agents today runs 30,000 tomorrow. Without ownership of retirement, versioning, and access, the estate becomes technical debt with decision-making authority. Governance of this kind belongs to the CEO and the board, not to a working group.
The org chart thins in the middle
Kutcher's forecast for the future organization: flatter structures, more workers exercising judgment, fewer managers coordinating them. Fast learners and people willing to challenge advance. Roles built on routine coordination compress.
Two consequences follow for succession planning:
- Fluency now flows upward. Junior employees frequently hold more practical command of the technology than the executives directing them, which makes reverse mentorship a structural requirement rather than a gesture.
- Apprenticeship still has to be funded. Cutting the early-career rungs solves a one-year cost problem and creates a ten-year judgment problem. No enterprise hires its way into a bench of senior operators with no place to grow them.
What this changes about executive hiring
The CEO agenda described above needs people who have done the work before, and the market for them is thin.
- Ownership moves out of the lab. Business units take accountability for AI-enabled problem solving instead of routing requests to a central team, which raises the bar on every P&L leader in the company.
- New titles carry real P&L weight. Chief AI Officer, Head of AI Transformation, Head of Physical AI and Robotics. Christian & Timbers placed Dr. Hui Cheng as Head of Physical AI and Robotics precisely because the mandate now sits at the top table.
- Deployment talent is the binding constraint. Our six-month study of the forward deployed engineer market identified roughly 2,000 elite FDEs in the United States against projected demand growth of 2,100%. Enterprises competing for the people who translate models into working workflows are competing against frontier labs for the same few hundred names.
- Compensation has repriced faster than most committees have modeled. Our 2026 Corporate AI Compensation Study benchmarks base, bonus, and equity for AI leadership roles across public companies, and the spread between market and stale internal bands is where most searches stall.
Seven questions for the board
- Who personally owns AI governance, and does the name appear on the CEO's scorecard?
- Which end-to-end workflows have been redesigned, as opposed to augmented?
- What share of EBIT do we attribute to AI, and by what method?
- How many senior leaders have built something with the technology in the past 90 days?
- What is the ambition, and does the current process make the ambition reachable?
- How do we retire, audit, and version agents already in production?
- Do our compensation bands match the market for the roles named in the plan?
A board unable to answer the first four is governing a technology program, not a transformation.
The evidence across a decade of survey data converges on one point. Companies capturing financial value from AI are led by CEOs who use the technology, redesign the work, own the governance, and hire ahead of the plan. The other 94% have a value problem wearing the costume of a technology problem.
Christian & Timbers builds the leadership teams behind the first group. Our 2026 Corporate AI Compensation Study gives boards and CHROs the benchmarks needed to compete for them.