MIT just released research that should make every CEO stop and pay attention.

According to The GenAI Divide: State of AI in Business 2025 - based on analysis of 300 public AI deployments, 150 executive interviews, and surveys of 350 employees - only 5% of enterprise AI pilots achieve measurable business results. The other 95% fail, delivering zero return despite an estimated $30–40 billion in global enterprise investment.

Read that again. Ninety-five percent.

And yet the boardroom conversation in most organizations right now is: "How do we move faster on AI?"

That's the wrong question.

When I work with leadership teams on transformation, I often ask them to picture their organization as a building. Solid foundation. Load-bearing walls. A roof that keeps the weather out. Everything is designed to be stable, permanent, predictable. You know where the exits are. You know who sits where.

Then I ask them: what happens when you try to add a new wing? Or move rooms around? Take down a wall?

You hire architects. You bring in contractors. You vacate the space, block off sections, deal with dust and noise and delays. People work around the construction for months. Some leave because the disruption is too much. And when it's finally done, you hope you don't ever have to do it again...until the next renovation is needed.

That's exactly how most organizations are approaching AI right now. They're treating it like a renovation project. A big, expensive, disruptive project with a finish line. Get the pilot done, prove the ROI in six months, present the results to the board, move on.

But AI isn't a renovation. It's not something you install. It's an ongoing shift in how work gets done, and it will keep shifting. The models that exist today will look primitive in two years. The use cases that matter in 2026 will be different in 2028. If your organization needs to launch a new initiative every time the technology evolves, you'll be in permanent renovation mode, and your people will feel it.

Here's what the MIT research actually tells us, underneath the headline number.

The study found that tools fail in enterprises not because of the quality of the AI itself, but because of what the lead researcher called "the learning gap" — tools and organizations both failing to adapt to each other. Generic tools stall because they can't learn from or integrate with real workflows. But more importantly, organizations stall because they haven't built the internal capacity to absorb, apply, and evolve with new technology.

The 5% that succeed? They're not smarter companies. They're not spending more. They're doing something structurally different: they're deploying AI into organizations that already know how to adapt. They have line managers empowered to drive adoption. They hire and develop resilient employees. They select tools that integrate deeply into existing workflows. They measure the right things. They don't depend on AI to hand down the answers.

In other words, the 5% aren't running better AI projects. They're operating more like ecosystems than buildings. Nature continuously adapts. A river flows downstream, carving its path, water levels rising and falling, snaking through valleys. It doesn't stop at an obstacle. It navigates around, over, through.

Bain & Company's 2024 research adds another layer to this. Their analysis of more than 24,000 transformation initiatives found that 88% of business transformations fail to achieve their original ambitions. Not just AI transformations - all transformations. And the cause isn't strategy. It isn't even talent, though that's what Bain focuses on. It's that organizations are running transformation as a series of discrete, heroic efforts rather than a continuous organizational capability.

Think about what that means in practice. Your organization just finished a major ERP implementation. Or a restructuring. Or a culture change initiative. Or an Agile transformation. And now you're being asked to "transform for AI." Each one is its own project, its own consultants, its own budget, its own change fatigue.

That's renovation after renovation after renovation. No wonder 88% fall short.

The organizations I work with that are actually navigating this well share one characteristic: they've stopped treating transformation as something that happens to them and started building the capacity to adapt as part of how they operate. It's not a project. It's not an initiative. It's a natural part of their DNA - the kind that's cultivated and nurtured over time.

This is Adaptive Capability. Not change management. Not agile methodology. Not another framework to implement and then forget. The actual organizational muscle to sense what's changing, integrate it, and keep moving without the need for expensive, messy, and risky renovation.

When that muscle exists, AI adoption looks different. It's not a 95% failure rate because the organization already knows how to integrate new ways of working. Line managers don't need to wait for a central directive because they have the authority and the skills to lead adoption in their own areas. Resistance isn't a blocker because psychological safety and learning culture are already embedded. The tools get used because the organization has built the habit of adapting to tools.

Here's what I'd challenge you to sit with: if your organization launched an AI pilot today and it failed, what or who would you blame?

Most executives would say the technology wasn't ready. Or the vendor oversold. Or the timing was off. Or the team didn't have the right skills. Or employees are just resistant.

Very few would say: "We don't have the organizational infrastructure to absorb this kind of change."

That's the answer that actually fits the data. And it's the one worth taking seriously.

The MIT study is getting framed as a cautionary tale about AI hype. I'd argue it's a much older story with a new headline. We've been here before with ERP. With Agile. With digital transformation. The technology changes. The failure mechanism stays the same.

What changes when you cultivate adaptive capability isn't that change becomes easy. It's that you stop needing a renovation every time the world moves.

If this is resonating with what you're seeing in your organization, I run an Adaptive Capability Diagnostic — a comprehensive assessment of your organization's adaptability maturity that gives leadership teams a clear picture of where the gaps are, they those gaps exist, and what to do about them.