AI Doesn’t Need More People Using It. It Needs People Thinking Differently.

One of the biggest mistakes leadership teams are making with AI is asking people to use AI without changing the way they work.

It looks sensible on the surface. Give everyone access. Run the training sessions. Track adoption. Count the prompts. Celebrate the dashboard turning green.

The board is told that AI adoption is increasing and progress is being made. Yet the business impact remains stubbornly modest. That should make leadership teams uncomfortable.

The problem may not be that people aren't using AI. It may be that they are using it inside an organisation that hasn't changed.

THE ADOPTION TRAP

There is an understandable temptation to measure what is easy to measure.

How many employees have access? How many are active users? How many prompts have been submitted? How many AI tools have been deployed? How many pilots are underway?

These numbers can tell you whether something is happening. They cannot tell you whether the organisation is creating value.

Recent research points in the same direction. MIT's NANDA research found that most enterprise GenAI deployments it examined were not producing measurable business impact. Research from the NBER has also found a significant gap between reported AI use and measurable productivity improvement. McKinsey's research similarly points towards a distinction between individual productivity gains and broader enterprise-level impact.

The precise numbers matter less than the pattern.

People can become more productive at individual tasks without the organisation becoming more productive overall.

That is the problem.

If someone can produce a report in half the time but still has to attend the same meeting, complete the same approval process, update the same systems and wait for the same decision, the organisation has not necessarily transformed.

It has simply made one part of the process faster.

AI CAN CHANGE THE TASK WITHOUT CHANGING THE JOB

This is where the human experience becomes important.

Imagine someone whose job involves reviewing information, preparing recommendations and passing them to another person for approval.

AI can now summarise the information.

It can draft the recommendation. It can identify anomalies. It may even suggest what should happen next.

The employee becomes faster. The organisation still requires the same approval. The same handoff remains. The same meeting takes place. The same reporting is completed. The same manager signs it off. The same hierarchy remains.

The person has been given a more powerful tool.

The job has not fundamentally changed.

This is one of the reasons AI programmes can feel transformative to leadership teams while feeling surprisingly ordinary to the people doing the work.

The technology has changed. Their working day has not. It has actually become more complicated.

They now have to check the AI output, correct it, explain it, transfer it into another system and remain accountable for the final result.

AI has been added to the workflow.

The workflow itself has survived untouched.

FASTER IS NOT THE SAME AS TRANSFORMED

There is a simple leadership trap here.

If an inefficient process takes ten hours and AI reduces one part of it by 50%, the organisation may celebrate a five-hour saving.

If another part of the process still takes eight hours, the overall outcome may barely change.

The bottleneck has simply moved. Review becomes the constraint. Coordination becomes the constraint. Approval becomes the constraint. Decision rights remain unchanged.

The organisation has accelerated activity without redesigning the system that turns activity into value.

That is why some AI programmes feel remarkably busy while producing remarkably little.

Faster execution of an unchanged workflow does not necessarily create transformation.

It simply produces yesterday's outcome more quickly.

THE QUESTION LEADERS NEED TO ASK

There is one question I would put in front of every leadership team reviewing an important workflow: If AI can now perform part of this work, why are we still organising the work in the same way?

That question changes the conversation. It moves the discussion away from software and towards organisational design.

If AI can perform the first stage of a process, do we still need the same handoff? If AI can analyse the information, should the person previously responsible for collecting it still spend most of their time doing that? If an agent can monitor something continuously, does the organisation still need the same reporting meeting every week? If AI can prepare a recommendation, should the decision still pass through three levels of approval? If AI can handle the routine part of a role, what should the person now spend their time doing?

These are not technology questions. They are leadership questions.

THE SACRED COWS

Every organisation has processes that survive long after the reason for creating them has disappeared.

The approval that once protected the business but now slows it down. The meeting that exists because nobody has challenged its purpose. The report that nobody reads but somebody still produces. The handoff created when systems could not talk to one another. The manager whose role grew around coordinating work that technology can now coordinate differently.

AI exposes these things. Technology alone will not remove them. Someone has to ask why they still exist. That is where leadership becomes uncomfortable.

Redesigning work means challenging structures, responsibilities and sometimes people's understanding of their own roles.

It is much easier to buy another licence than to tell a senior manager that three layers of approval are no longer necessary. It is easier to run an AI workshop than to redesign a department. It is easier to announce an AI strategy than to decide what the organisation should stop doing.

Transformation becomes real when something changes that people can actually feel.

WHAT THE HIGH PERFORMERS ARE DOING DIFFERENTLY

The research increasingly points towards workflow redesign as one of the characteristics separating organisations that capture meaningful value from those that simply adopt the technology.

McKinsey's research, for example, has found a strong relationship between workflow redesign and organisations reporting meaningful AI impact. That matters because it changes the conversation.

The question is no longer: “How do we get more people using AI?”

It becomes: “How should this work now be organised because AI exists?”

That is a much more difficult question. It is also a much more valuable one.

THE PEOPLE DOING THE WORK ALREADY KNOW

There is another opportunity that leadership teams often overlook. The people closest to the work can usually see where the friction is.

They know which steps are duplicated. They know which reports nobody reads. They know where customers wait. They know which approvals add little value. They know where people create workarounds because the official process doesn't work.

Increasingly, many of them are already experimenting with AI themselves.

The problem is that they may not believe they have permission to challenge the process. They have been told to use AI. 

They have not necessarily been told: “You are allowed to question the way we work.”

That distinction matters.

If leaders want AI to create organisational change, they need to give people permission to challenge the workflow, not simply permission to use the tool.

They need to create enough psychological safety for someone to say: “We don't need to make this step faster. We need to stop doing it.”

That may be one of the most valuable sentences an organisation hears during an AI transformation.

AI CHANGES CAPABILITY. CAPABILITY CHANGES WORK.

This is where the implications become bigger than productivity.

If AI changes what people can do, it changes the shape of work. If the shape of work changes, roles may need to change. If roles change, decision rights may need to change. If decision rights change, management structures may need to change.

If management structures change, the organisation itself may need to change.

That is the chain many AI strategies overlook.

AI changes capability. Capability changes work. Work changes decisions. Decisions can change organisations.

This is also why AI transformation cannot be delegated entirely to the technology function.

The CIO can implement the technology.

The CTO can build the capability.

The CISO can manage the risk.

The HR function can support the people implications.

Someone in the leadership team has to decide how the organisation should work differently because the technology now exists.

That is an organisational decision.

WHAT LEADERS ACTUALLY NEED TO DO

For significant AI initiatives, I would ask leadership teams to focus on five things.

1. Measure changed outcomes, not just adoption.

Look at cycle time, decision quality, error rates, customer outcomes, capacity released and financial impact. Ask what is different, not simply who is using the technology.

2. Give people permission to challenge the process.

Don't just train employees to use AI. Ask them where the existing workflow no longer makes sense.

3. Redesign roles and workflows alongside the technology.

Don't wait until the technology has been embedded for a year before asking how the organisation should adapt to it.

4. Be explicit about what humans still own.

Automation does not remove accountability.

If AI performs part of a decision, someone still needs to understand the decision, challenge the output when necessary and own the consequence.

5. Be prepared to stop doing things.

This is probably the hardest one.

Every successful transformation creates something new.

The best ones also remove something old.

THE QUESTIONS I WOULD PUT TO THE LEADERSHIP TEAM

Before celebrating AI adoption, ask:

What work has actually disappeared? What work has changed? What decisions are now made differently? What capacity has been released? What should our people now spend more time doing?

Which approval, meeting, report or handoff no longer makes sense?

The most uncomfortable question: What are we still doing simply because that is how we have always done it?

If the answers are difficult, that doesn't mean the AI programme is failing. It may mean the real transformation work is finally beginning.

THE REAL COMPETITIVE DIFFERENCE

The organisations that win with AI will not necessarily have the most sophisticated models.

They will not necessarily have the highest adoption rates. They may not even have the biggest AI budgets. They will be the organisations whose leaders are prepared to change how work is actually done.

That requires more than technology. It requires curiosity. It requires judgement. It requires organisational courage. It requires leaders who understand that giving people a better tool is not the same as giving them a better way of working.

The competitive advantage will come when the organisation begins to ask a different question.

Not: “How do we get our people to use AI?”

But: “Given what our people and our technology can now do, how should we work?”

That is a much bigger question.

It is the question that will separate AI adoption from AI transformation.

AI doesn't need more people using it inside yesterday's organisation. It needs people, and leaders, willing to rethink the organisation itself.