The Number That Should Keep Every CEO Up at Night
Here is a stat that will make you uncomfortable: only 16% of CEOs think their company’s infrastructure is inadequate for AI. Among IT leaders in the same companies? That number jumps to 42%.
That is a 26-point gap between the people making the AI investment decisions and the people who actually have to make AI work. This data comes from Kyndryl’s 2025 People Readiness Report, and it is one of the clearest signs that most CEO AI strategies are built on blind spots, not data.
If you are a CEO planning your company’s AI future, the biggest risk is not the technology. It is what you do not know about your own organization.
The CEO vs. CTO Perception Gap
The disconnect goes deeper than infrastructure. When Kyndryl surveyed business leaders and tech leaders at the same organizations, the gaps showed up everywhere.
On employee readiness: 45% of CEOs believe their employees embrace AI. Among tech leaders, 73% say the same. CEOs are almost 30 points more pessimistic about their own teams than the people who work with those teams every day.
On skills: 76% of leaders say reskilling is a priority over hiring. Sounds good on paper. But fewer than 40% actually run reskilling programs. The intention is there. The execution is not.
On infrastructure: CEOs see modern systems and cloud deployments. IT leaders see legacy databases, disconnected APIs, and data sitting in 12 different spreadsheets that nobody owns.
This is not a technology problem. It is an information problem. CEOs are making strategic decisions with incomplete data about their own companies.
Why This Gap Exists (And Why It’s Dangerous)
CEOs operate at altitude. That is the job. You set direction, allocate budget, and trust your teams to flag problems. The board wants an AI strategy. Competitors are announcing AI initiatives. The pressure to move fast is real.
But AI readiness is different from most strategic initiatives for three reasons.
1. AI exposes hidden weaknesses. Your ERP might work fine for daily operations. But when you try to feed its data into an AI model, you discover that 80% of the effort is data preparation, not model building. The “working” systems are actually full of gaps that only surface when AI needs them.
2. The feedback loop is broken. IT teams know the infrastructure is not ready. But translating “our data governance is inadequate” into language that creates CEO urgency is hard. Technical teams speak in system dependencies and data pipeline constraints. CEOs speak in revenue, market share, and competitive advantage. So the message gets softened. “We might need some upgrades” instead of “we are 18 months away from being able to run production AI.”
3. Vendor demos lie. Every AI vendor shows a polished demo with clean data. Your data is not clean. Your systems are not integrated. The gap between what the demo shows and what your company can actually do is where AI projects go to die. And they do: 80% of AI projects fail to reach meaningful production, according to RAND Corporation.
What a Real CEO AI Strategy Looks Like
A CEO AI strategy that actually works does not start with picking tools or running pilots. It starts with honest assessment. Here is what that means in practice.
Close the perception gap first. Sit down with your CTO or IT director and ask one question: “What are the top 5 things that would prevent us from deploying AI in production within 6 months?” Do not argue with the answers. Write them down. That list is your real starting point.
Audit your data, not your ambitions. Before you sign any AI vendor contract, know the state of your data. Is it centralized or scattered across departments? Is there governance, or does everyone maintain their own version of the truth? AI readiness has 7 dimensions, and data foundations carry the heaviest weight at 25%.
Fund the boring stuff. The unglamorous work of cleaning data, connecting systems, and documenting processes is what makes AI possible. It is not exciting in a board presentation. Nobody wins an innovation award for data governance. But skipping it is why 74% of companies have not seen real value from their AI investments, according to BCG. The companies that succeed with AI spend 60% of their budget on data foundations before they write a single line of model code.
Match reskilling words with reskilling budgets. If 76% of your leadership team says reskilling matters, but you are spending less than 2% of your AI budget on training, the strategy is a slide deck, not a plan. Your people need to understand what AI can and cannot do before they can use it effectively. That means hands-on workshops, not a 30-minute webinar. It means giving teams time to experiment, not just telling them to “use AI more.”
Build a common language between the boardroom and IT. The perception gap thrives on ambiguity. When the CEO says “we need AI,” the CTO hears a six-month infrastructure project. When the CTO says “the data isn’t ready,” the CEO hears resistance to change. Create a shared assessment framework with specific metrics that both sides agree on. Numbers cut through assumptions.
The 5-Question CEO Self-Assessment
Before your next board meeting about AI, answer these honestly:
1. Can your IT team deploy a working AI prototype within 90 days using your existing data? (If you don’t know the answer, that is the answer.)
2. When was the last time you asked your CTO to rate your AI readiness on a scale of 1 to 10? (Not your ambition. Your readiness.)
3. What percentage of your business data is centralized, documented, and accessible via API? (If the answer is “I’m not sure,” it is probably below 30%.)
4. Do you have an active reskilling program running right now, or is it still in the planning phase?
5. Has anyone on your team completed an independent AI readiness assessment, or are you relying on vendor assessments? (Vendors will always tell you that you are ready. That is how they sell.)
If you answered “no” or “I don’t know” to three or more of these, you are in the same position as most CEOs. You are not behind. You are normal. But normal is not where you want to stay, because your competitors are asking these same questions right now.
The good news: awareness of the gap is the first step to closing it. And closing it does not require a massive transformation budget. It requires honest information.
Stop Guessing. Start Measuring.
The CEOs who will succeed with AI in 2026 and beyond are not the ones with the biggest budgets or the flashiest pilots. They are the ones who closed the perception gap between the boardroom and the server room.
That means getting honest data about where your company actually stands. Not where you think it stands. Not where your last consultant said it stands. Not where the vendor demo made it look like it stands. Where it actually stands today, across all seven dimensions: strategy, data, technology, people, culture, processes, and governance.
The perception gap is not a failure of intelligence. It is a failure of information flow. Fix the information, and the strategy fixes itself.
At Proxima, we run structured AI readiness assessments for Greek and European businesses. No vendor bias. No sales pitch disguised as an audit. Just a clear picture of where you are, where the gaps are, and what to do first.
Because the most expensive AI strategy is the one built on assumptions.
Related Articles
Keep Reading
- The 7 Dimensions of AI Readiness
- Why 80% of AI Projects Fail
- How to Write an AI Strategy
- AI Implementation Roadmap: 90 Days
Need help putting this into practice? Our Consulting Services or Let’s Talk.