AI-ready? Prove it.

Opening

What does it take to be an AI-ready leader? Two recent studies from Boston Consulting Group converge on an uncomfortable answer. 36% of executives believe their senior leaders are ready to lead AI transformation. 25% of employees feel their leaders support their use of AI.

Two different studies. The same blind spot. Companies are measuring AI readiness through employee training and adoption rates, while underestimating leadership's own responsibility to make the organisation AI-ready.

Employees can be AI-literate without the company being AI-ready.

Employees can be trained on the tools. They can learn the prompts, run the workflows, and improve their individual productivity. None of that, on its own, makes the organisation AI-ready.

Employees cannot, by themselves, change the operating model. They cannot eliminate the bottlenecks that stall AI from reaching decisions. They cannot redefine how decisions get made, where accountability sits, or how functions coordinate across the business. Those are leadership responsibilities.

And no amount of workforce training compensates for leadership not doing them.

What AI-ready leadership actually owns.

AI-ready leadership owns three things employees cannot deliver on their own.

The operating model. Where AI enters the workflow, which decisions it supports or replaces, what evidence is required for sign-off, and which human roles remain accountable at each step. This is a design job — and it sits with leadership.

The information flow. AI adoption generates real-time signal from the people closest to the work — reliability issues, unexpected risks, opportunities for improvement. Leaders must ensure that information reaches the people who make decisions about scale, investment, and risk — rather than staying trapped in the team that noticed it.

The culture of speaking up. Teams working directly with AI see what is not working first. If they are not encouraged — and expected — to raise it, the organisation learns slowly, if at all. This is not a soft skill. It is a governance condition for AI that compounds.

What two-way adoption looks like.

At Amazon Web Services, I saw the value of treating AI adoption as a two-way process. Teams were encouraged to raise issues, share practical feedback, and suggest improvements. The lessons from everyday AI use informed broader decisions, improved adoption at scale, and sharpened the identification of opportunities for business value.

The direction travelled both ways. Leadership set the strategic frame; teams brought back the operational truth. Neither side could have delivered AI value alone.

That two-way discipline is what most organisations now claim to want and few organisations have actually built. It does not emerge from training programmes. It emerges from leadership designing the feedback loops — and then rewarding the people who use them.

What the Board should actually be asking.

Training completion, adoption rates, and the number of use cases underway are the metrics most Boards see today. They are the easy numbers to collect. They look like progress. And they obscure the questions that matter.

The visual below compares what most Boards see on AI dashboards today against the questions that would actually reveal whether the organisation is becoming AI-ready.

Where this leaves leaders.

Employees are expected to change how they work with AI. That expectation is reasonable, and it is now widely held.

Leaders must be willing to change how the company works, how decisions are challenged, and how feedback moves. That expectation is less widely held. And it is the one that separates organisations where AI reaches the P&L from organisations where it does not.

This is the work we do with the Boards and executive teams we advise — moving AI from training completion to structural readiness, with the leadership discipline the role now demands.