AI Doesn't Transform Businesses. Leaders Do.

IN ONE MINUTE

AI doesn't transform a business on its own. Leaders do, by setting a clear purpose for what the business needs to become, and using AI to get there. McKinsey's study of 20 companies with successful AI-enabled transformations found an average EBITDA improvement of about 20%, and around $3 of incremental EBITDA for every $1 invested, with a C-suite that owned the change as the most important driver. A simple test shows whether a programme is ready: can you finish the sentence "Because of AI, we will now be able to ___, which changes our business model by ___"? If not, it is a technology programme, not a transformation.

THE OBJECTIVE

I have spent several years leading AI and digital transformation work, and I have come to believe the phrase "AI transformation" is aimed at the wrong noun.

AI doesn't transform anything. It sits there, extraordinarily capable, waiting. What transforms a business is a leader who decides what that business needs to become, and then uses AI as a tool to get there. In my experience, that single distinction separates the programmes that deliver real returns from the ones that quietly fade away.

THE CONFESSION: CHASING THE SHINY OBJECT

Early in my AI journey, I fell straight into the trap. I chased the technology.

New model? Deploy it. New capability? Build a pilot. Each one felt like progress. The activity reports looked impressive and the dashboards were green.

But the business itself, how people made decisions and how value moved through the organisation, barely changed. I was leading a technology programme and calling it a transformation. They are not the same thing, and no amount of budget closes the gap between them.

Everything changed when we stopped asking "What can this AI do?" and started asking "What should this business become, and where does AI actually help us get there?" The second question is much harder. It is also the only one that matters.

WHAT THE RESEARCH SHOWS

McKinsey's recent study, drawn from the second edition of Rewired, looked at 20 large companies that have created significant value from AI-enabled transformation:

• Average EBITDA improvement of about 20% after three years.
• Around $3 of incremental EBITDA for every $1 of one-time investment.
• Cash positive within one to two years on average, by sequencing early wins that paid for the journey.
• Two-thirds focused on three business domains or fewer.

The authors are clear on the most important driver: a C-suite that understands AI and treats the transformation as its own work, together, rather than delegating it.

BCG reaches a similar view from a different angle. Its research suggests around 70% of AI's value comes from changing how people work and how processes run, with the rest from algorithms and technology.

THE DOCTRINE: STRATEGY FIRST, TOOL SECOND

Here is the belief I now hold with very little room for negotiation: start with purpose. Before any model, platform or pilot, be clear what the business must become and why it matters to customers, people and results. AI is either serving that purpose, or it is corporate theatre. Purpose is what turns a technology programme into a transformation.

The companies generating real returns rarely talk about "AI transformation" internally. They talk about faster decisions, better customer experiences, leaner operations and new revenue lines. AI shows up quietly, in service of an outcome someone already believed in.

The programmes that struggle tend to talk about the technology itself, models, platforms and roadmaps, but can't explain in one simple sentence what business outcome any of it serves.

THE FRAMEWORK: THE ONE-SENTENCE TEST

If you can't finish this sentence, your transformation isn't ready:

"Because of AI, we will now be able to ___, which changes our business model by ___."

Not "we'll have better tools". Not "we'll be more efficient". The answer should point to a real business shift. These signals help tell the two apart:

	Technology programme	Business transformation
The starting question	What can this AI do?	What should this business become?
How progress is reported	Pilots launched, users, usage, dashboards	Decisions, customers, costs, revenue
Who leads it	The technology team, with the business consulted	The business leadership, with technology as a close partner
Where the effort goes	Building and deploying tools	Redesigning how work, decisions and value flow
What "done" looks like	The tool is live	The business model or operating model has changed

WHAT THIS MEANS FOR CEOS AND EXECUTIVE TEAMS

Own the transformation together. McKinsey's most important driver was a C-suite that treated the change as its own work. Delegating AI to a single function signals that it is a technology project.

Start with the business problem. Not "where can we use AI?" but "what would move our results most, and could AI help?" Ask every proposal to pass the one-sentence test before it is funded.

Narrow the focus. Two-thirds of McKinsey's successful companies concentrated on three domains or fewer. A few high-value areas, redesigned end to end, beat dozens of disconnected pilots.

WHAT THIS MEANS FOR P&L AND FUNCTIONAL LEADERS

If you lead a function, a P&L or an enterprise, the most valuable AI question right now isn't "What should we build?" It is: "What are we trying to become, and does that require AI, or does it just sound better on a slide with the word attached?"

Look for the non-technical decisions. Some of the best "AI decisions" I have been part of had nothing to do with code. They were decisions to simplify a process, remove a layer of approval, or trust a team with more autonomy. AI made them possible, but a leader had to choose to act on them.

Review progress in business terms. Stay close to the work, remove blockers, and change course when the evidence says so. Plans rarely survive first contact unchanged, and that is normal.

WHAT THIS MEANS FOR TECHNOLOGY AND TRANSFORMATION TEAMS

Translate capability into outcomes. Present every initiative in terms of the business result it serves, not the model or platform behind it.

Measure what the business measures. Usage and adoption are early signals, not results. Link each initiative to a decision, customer or financial outcome.

Be a partner, not a supplier. The strongest programmes pair a senior business owner with technology leadership from day one.

The technology is quickly becoming a commodity. The judgement of what to become with it is not. That judgement remains a leadership act.

THE LAST WORD

AI changes what's possible. Leaders decide what's next.

What is the one sentence your organisation would need to finish, "Because of AI, we can now ___", if someone asked you today?

—

Until next week,

Artificial Ian

Tell your network: if your leadership team is setting next year's priorities, the one-sentence test is a useful way to open the first session.

→ Found this useful? Forward it to a leader still chasing the tools instead of the outcome.

→ Subscribe to The AI Executive Briefing, weekly in your LinkedIn inbox.

Sources: McKinsey Quarterly, "The new management playbook for AI" (Aug 2026), drawing on Rewired, second edition; BCG, "What Separates CEOs Who Scale AI From Those Who Don't?" (Jul 2026).