Most leadership teams I talk to have at least one leader who is confident about their artificial intelligence (AI) readiness. The decks are sharp, the pilots are running, the vendor conversations are progressing.
One pattern I have noticed: confidence rarely extends more than a level or two down in the organization. One company I spoke with recently was confident their email communication to employees about AI had landed as intended. When we looked more closely, the communication was being pushed down, but no one was asking for input, and little was coming back up from the people actually doing the work.
The gap between what leaders believe about their AI readiness and what their own people believe is measurable. It is directional. And it is the single most predictable reason mid-market AI initiatives stall in year two.
What follows is about that gap: where it comes from, why no one inside the company surfaces it, and what to do about it before the strategy gets built on top of a picture of the company that is not quite true.
The gap is measurable, and directional
Korn Ferry's Workforce 2025 research, a study of 15,000 global employees, found that 78% of leaders believe they have AI figured out for their organizations. 39% of employees agree.1
That is a 39-point delta inside the same companies.
Translated into mid-market terms: for a 150-person company, your leadership team is confident about something that roughly three in five of your people do not believe is true. The direction of the gap is what makes it expensive. Leaders are systematically more optimistic than employees about the same organizational question. Optimism in the boardroom, skepticism in the trenches.
The downstream evidence is already in the data. DDN's 2026 State of AI Infrastructure Report, a survey of 600 U.S. IT and business leaders, found that 54% of organizations have delayed or canceled AI initiatives in the past two years.2 Sixty-five percent say their AI environments are now too complex to manage.
These are not, at root, technology failures. They are almost always the confidence gap showing up eighteen months later, after the strategy has been committed and the execution has run into a reality that no one could see when the decisions were made.
Why the gap forms, and why no one inside the company surfaces it
Leaders see the pilot. Employees see the implementation.
The pilot is the version of AI that was built in a controlled environment, on a clean dataset, with an enthusiastic early-adopter team. It works. Executives see it, it confirms their optimism, and the next conversation is about scale.
Employees see something different. They see the workflow friction the pilot did not surface. They see the data problems the vendor demo did not touch. They see the colleagues who have quietly stopped using the tool because it slows them down. They see the colleagues who use it but are not telling anyone, because they are not sure whether transparency will be rewarded or punished.
And the people best positioned to close that information gap (the directors and senior managers who sit between the boardroom and the work) have often been trained not to contradict the CEO in a room where the CEO is optimistic. The structural silence is not a failure of courage. It is a rational response to an organizational system that rewards alignment and penalizes dissent.
What results is something quieter and more expensive than misalignment. A CEO is making strategic bets on a picture of their own company that is missing several of the most important facts. The gap is invisible to the person whose decisions depend on it most.
What the gap costs
Every AI transformation runs on two engines: the financial engine, where the return on investment (ROI) shows up in savings, productivity, and new revenue, and the human engine, where it shows up in trust, morale, and the organization's capacity to absorb change. A strategy that only measures the first will underperform predictably when the second is stalled.
The confidence gap is a human-engine signal the board does not track. But it is a leading indicator for the failure modes that eventually do show up in the profit-and-loss statement.
Some of those failure modes are already visible in the research. Forrester has reported that 55% of employers regret laying off workers for AI.3 Gartner forecasts that by 2027, half of companies that attributed customer service headcount reductions to AI will rehire staff to perform similar functions, often under different job titles.4 Both are patterns of decisions made with confidence that later turned out to have been made without the reality check.
The cost of moving without the reality check is not theoretical. It is the cost of the strategy that stalls, the layoff that reverses, the initiative that gets delayed twice and then quietly shelved.
Clarity before strategy
I wrote in an earlier piece, You Know AI Matters. That's Not the Problem., that the stuck AI initiative is usually not a knowledge problem. Executives already know AI matters. They are stuck on what to do with that knowledge, and the real constraint is almost always clarity, not information.
The organizations I have watched build durable AI strategies start in a different place than a technology decision. They start with an honest diagnostic of where they actually are.
That diagnostic has to look at more than one dimension. AI readiness is not a single number. It has five load-bearing components, each of which must be understood independently before any of them informs a strategy:
- Technology Infrastructure: what your systems actually support, independent of vendor claims
- Data Maturity: whether your data is organized, accessible, and trustworthy enough to feed the use cases you are considering
- Culture & Change Readiness: whether your organization can absorb transformation at the pace the strategy requires
- Competitive Position & Strategy: where AI should actually create measurable impact for your business, not where it is most fashionable
- Workforce Capabilities: what your people can do now, what they will need to learn, and what roles will shift
Most leadership teams have a strong read on one or two of these, often strategy, sometimes workforce. The rest are a blur of assumption, vendor narrative, and the confidence this post opened with. The strategy ends up built on the parts of the picture leaders know well, while absorbing the risk of the parts they do not — without pricing it in.
A multi-dimensional diagnostic replaces assumption with measurement across each pillar. The confidence gap then becomes visible not as an abstract belief problem but as a specific, actionable finding (our data maturity is further behind than we have been treating it; our culture readiness is less prepared for this pace of change than the strategy assumes) that the leadership team can make decisions on.
Clarity before strategy. The organizations that skip this step do not move faster. They just discover the gaps later, when the cost of finding them is higher.
What this looks like for a 150-person company
Three concrete moves a leadership team can make this quarter, before committing to an AI strategy:
- Run the five-pillar diagnostic before the vendor conversations get serious. Technology, data, culture, competitive position, workforce. Each one examined independently, by someone who is not selling anything downstream. The dimensions that matter most to your strategy are rarely the dimensions you are naturally most curious about.
- Compare what you assumed with what the diagnostic surfaced. Name the specific pillars where your read was further from reality than you expected. That delta is the signal — and the pillars where the delta is largest are the pillars where execution will stall first.
- Design the first 90 days of the strategy around the highest-leverage gap — not around whichever AI initiative is most shovel-ready. If data maturity is the weak link, address data maturity first. If culture readiness is the weak link, sequence the change plan accordingly.
Re-run the diagnostic every six months. AI readiness moves as the technology moves, as the competitive landscape moves, and as your people's experience of AI at work accumulates.
If you are about to build an AI strategy, start here
If you are preparing to invest meaningfully in AI (a strategy engagement, a platform decision, a workforce move), the single most useful thing you can do first is replace your assumptions with a measured picture of where your organization actually sits.
The AI Readiness Assessment is designed to give you that picture across the five pillars that determine whether a strategy will hold: technology infrastructure, data maturity, culture and change readiness, competitive position and strategy, and workforce capabilities. It replaces leadership assumption with measured readiness, one pillar at a time. It is an intelligence tool, not a sales gate. The strategy comes after.
For teams that want to go further — to unpack the specific confidence gap this post opened with, the one between what a leadership team believes and what their own people see — that work lives in a companion workshop format that builds on the diagnostic findings. More on how the two fit together on the services page.
Most leadership teams that sit down honestly with their own readiness picture discover two things. First, they have more to work with than they thought. Second, they are not as aligned as they assumed. Both of those discoveries are useful. Either one, if surfaced early enough, changes the shape of what comes next.
References
- Korn Ferry Institute. "How to Spot an AI-Ready Leader." Drawing on Korn Ferry's Workforce 2025 research (15,000 global employees). Accessed April 2026. https://www.kornferry.com/insights/featured-topics/gen-ai-in-the-workplace/how-to-spot-an-ai-ready-leader
- DDN. 2026 State of AI Infrastructure Report. Survey of 600 U.S. IT and business leaders conducted by Vanson Bourne, July–September 2025; published January 2026. https://www.ddn.com/2026-state-of-ai-infrastructure-report/
- Forrester research, cited in HR Executive, "The AI Layoff Trap: Why Half Will Be Quietly Rehired." https://hrexecutive.com/the-ai-layoff-trap-why-half-will-be-quietly-rehired/
- Gartner. "Gartner Predicts Half of Companies That Cut Customer Service Staff Due to AI Will Rehire by 2027." Press release, February 3, 2026. Note: the 50% figure is specific to customer service roles, per Gartner's survey of 321 customer service and support leaders. https://www.gartner.com/en/newsroom/press-releases/2026-02-03-gartner-predicts-half-of-companies-that-cut-customer-service-staff-due-to-ai-will-rehire-by-2027
This is Part 2 of the AI Transformation series.
More from the AI Transformation Series
- Part 1: You Know AI Matters. That's Not the Problem.
- Part 3: AI Readiness, Honestly Measured: The 5 Dimensions Most Leaders Don't Score Themselves On