AI adoption is often framed as a productivity story.
Faster workflows. Better decisions. More efficient teams. Stronger ROI.
But inside many organisations, AI is landing on people who are already stretched, squeezed and trying to keep up with a constant stream of change. Teams are dealing with restructures, shifting priorities, cost pressure and new expectations around what work should look like now that AI tools are widely available.
Leaders may see AI as a route to transformation. Employees may experience it as another layer of uncertainty.
What does this mean for my role? Which tools am I allowed to use? What happens to the parts of my job AI can now do? Am I expected to become more productive without any work being taken away? Will AI somehow replace what I do?
These questions are critical to answer because AI adoption happens inside teams made up of humans. And humans need clarity and reassurance before they can move with any level of confidence.
AI is increasing the pressure on already stretched teams
The pace of AI adoption is forcing organisations to make decisions faster than many people can absorb.
Henley Business School’s 2025 research captures this tension well. Its study found that 56% of full-time professionals felt optimistic about AI advancements, while 61% admitted they were overwhelmed by the pace of change. Henley described this as “FOBO”: feeling optimistic but overwhelmed.
That tells us something important. Many employees aren’t simply anti-AI.
They can see the potential. They can imagine how it might help them save time, improve quality or reduce repetitive work. But optimism fades quickly when adoption arrives without enough clarity, training or honest conversation.
The result is a familiar pattern: lots of AI activity, but uneven adoption and unclear impact.
Confusion slows adoption
When leaders talk about AI in broad terms, employees are left to fill in the gaps.
They hear phrases like “AI-first”, “unlock productivity” or “transform the business”, but they may not know what those phrases mean for their day-to-day work.
That fog creates several problems.
People may avoid using AI because they’re worried about getting it wrong. Others may use it on the side, creating shadow AI practices that leaders can’t see or learn from. Teams may duplicate pilots because nobody knows what’s already being tested. Managers may struggle to answer questions because they’re unclear themselves.
In the worst cases, AI becomes yet another source of fear.
That fear doesn’t always show up as visible resistance. It can look like hesitation, silence, passive compliance or endless requests for more information.
People wait for permission because the boundaries aren’t clear. They delay decisions because the priorities keep shifting. They hold back questions because they don’t want to sound slow, negative or difficult.
This is where clarity becomes critical.
Clarity is more than communication
Many organisations treat clarity as a communication job.
Announce the AI strategy. Share the new tools. Publish the policy. Run a town hall. Send the update. Maybe run one mandatory training session on Teams. Then hope for the best.
In my work with leaders, and in my book, I describe clarity as the discipline of focusing on what matters so people can pull in the same direction. For AI adoption, that means leaders need to make three things explicit: direction, priorities and guardrails.
Direction answers: why are we adopting AI, and what business problem are we trying to solve?
Priorities answer: where should teams focus first, and what can wait?
Guardrails answer: how should people use AI safely, responsibly and usefully?
Without those three things, AI adoption becomes noisy. People may be busy using tools, attending workshops and launching pilots, but that doesn’t mean the organisation is creating value.
McKinsey’s 2025 State of AI report points to a similar challenge. It describes AI use as increasingly widespread, but highlights that the move from adoption to scaled impact still depends on management practices across strategy, talent, operating model, technology, data, adoption and scaling.
In other words, the tools matter. So do the conditions around them.
Set the direction: name the job AI is here to do
“We need to use AI” gives people pressure, not direction.
If leaders want teams to engage properly, they need to explain the job AI is being asked to do.
Is it there to reduce repetitive admin? Improve customer response times? Help teams make sense of large amounts of information? Support better decision-making? Speed up content production? Improve forecasting? Free people up for higher-value work?
Each answer points to a different kind of adoption.
When the job is unclear, people make their own assumptions. Some will assume AI is mainly about headcount reduction. Others will assume it’s a productivity target in disguise. Some will experiment enthusiastically, but in ways that don’t connect to the organisation’s priorities.
Naming the job creates focus.
It also helps teams understand what good looks like. If AI is meant to reduce admin, leaders should ask whether admin is actually reducing. If it’s meant to improve decision-making, they should ask whether decisions are getting better, faster or more informed.
If it’s meant to improve customer experience, they should track whether customers are seeing the benefit.
Clear direction turns AI from a vague transformation ambition into something people can test, discuss and improve with confidence.
Pick the priorities: make the first use cases visible
AI can create a strange kind of overwhelm because it appears useful almost everywhere.
That’s exciting, but it can also scatter attention.
One team uses AI for content. Another uses it for coding. Another uses it for customer insight. Another uses it for admin. Suddenly, everyone is experimenting, but nobody knows which experiments matter most.
Deloitte’s 2026 State of AI in the Enterprise report shows how quickly access is expanding. It found that worker access to AI rose by 50% in 2025, while leaders are increasingly focused on ROI, safe and ethical practices, workforce readiness and moving from pilots to production.
That makes priorities even more important.
Leaders need to name the two or three AI use cases that matter most right now. Which teams are leading them? What problem are they solving? What can wait? What would make a use case worth scaling?
These questions help protect teams from performative adoption.
People don’t need another initiative that exists mainly to prove they’re keeping up. They need a clear sense of where AI can make work better, and where it will genuinely help the organisation deliver.
Priorities also reduce the fear that everything is changing at once.
If leaders say, “This quarter, we’re focusing on AI adoption on customer response times and internal knowledge search,” people know where to place their attention. They also know where not to spend their energy yet.
That clarity is a kindness when teams are stretched. The discipline of saying “not yet” matters as much as the ambition to move fast.
Clarify the guardrails: make safe action possible
Clarity also means being specific about boundaries.
Many employees are unsure what they can safely put into AI tools, which platforms are approved, when human review is required and where AI use would create legal, ethical or reputational risk.
If those boundaries are missing, people either avoid AI completely or use it in ways that create risk for the organisation.
PwC’s UK Workforce Hopes and Fears Survey 2025 suggests AI use is also uneven across organisations. Senior executives are far more likely to use AI agents and GenAI tools regularly, while many non-managers rarely or never use them at work.
Guardrails don’t have to be perfect from day one, but they do need to be clear enough for people to start.
They also need to be written in language people can understand. A 40-page policy may protect the organisation on paper, but it won’t help a manager decide whether their team can use AI to summarise customer feedback, draft internal comms or analyse meeting notes.
The aim is to give people enough structure to act with confidence.
That means leaders need to make the rules easy to find, easy to understand and easy to apply in real work. If people have to interpret the policy every time they want to use a tool, adoption will slow down.
Good guardrails don’t kill experimentation. They make safe experimentation possible and survivable if things don’t go to plan.
Clarity makes AI adoption more human
The pace of AI adoption is unlikely to slow down.
Leaders will keep facing pressure to find productivity gains, boost innovation and prove ROI. Teams will keep facing the emotional reality of changing work, shifting expectations and uncertainty about what comes next.
Clarity won’t remove all of that tension, but it can reduce unnecessary confusion.
When people understand the direction, they can connect AI adoption to a real business problem. When they understand the priorities, they can focus their effort instead of chasing every possible use case. When they understand the guardrails, they can act with more confidence.
AI adoption becomes more sustainable when leaders treat clarity as an operating discipline, not a one-off communication exercise.
Because people don’t need more noise about transformation. They need to know what matters, what’s expected, and where they have permission to move.
