Most organisations now have an AI plan of some kind. Fewer have a plan for the people who are meant to use it. The tools get bought, the licences get issued, and six months later a small group uses them well, a larger group has tried them once, and a fair number of people are using something else entirely on their own phones.
This guide covers why AI adoption is mostly a people challenge, what staff are worried about, where the gaps between leaders and their teams sit, and what leaders can do to bring people with them. It is part of our wider guide to change leadership.
Why is AI adoption a people problem?
AI adoption asks people to change how they work, what they are good at and, in some cases, what their job is. That is a change problem first and a technology problem second. The research is consistent on this point.
In a 2026 benchmark survey of senior data and AI leaders, 93% named culture and change management as the main barriers to AI adoption, far ahead of the technology itself (Lab Manager 2026). Kyndryl's 2026 survey of 1,100 senior leaders across eight countries found that only 23% of organisations say their workforce is fully ready for AI, and 79% of leaders expect the speed of AI to outpace their workforce, governance and operating models (Kyndryl 2026).
The people side is also where many organisations are doing the least. An Adecco Group study of 2,000 C-suite executives across 13 countries found that only 36% say their talent strategy clearly shows employees how AI will create opportunities for them, and only 39% are involving employees directly in redesigning jobs (Adecco Group 2026).
Why are Australians so wary of AI?
Leaders in Australia start with less goodwill than their counterparts in many other countries. The University of Melbourne and KPMG's 2025 study of trust in AI across 47 countries found that only 36% of Australians are willing to trust AI, and only 30% believe its benefits outweigh its risks, the lowest of any country surveyed. Only 24% had undertaken any AI-related training or education, and 77% agreed that AI regulation is necessary (Gillespie et al. 2025).
That scepticism shows up at work. People who already doubt the technology will read a vague announcement in the least generous way. A rollout that might work elsewhere with a launch email and a training video needs more explanation, more involvement and more patience here.
What are people worried about?
The obvious worry is losing their job, and leaders should take it seriously. In our experience it is rarely the only one, and often not the strongest. The deeper worries tend to be about:
- Competence. Capable, experienced people fear becoming beginners again, and being seen to struggle in front of colleagues.
- Identity. People whose value rests on knowing how things work worry that the knowledge will stop mattering.
- Fairness. Staff wonder who will capture the time AI saves, and whether it will simply mean more work for the same pay.
- Being watched. Some tools record and analyse how people work, and people want to know what will be done with that data.
- Quality. Many people have seen AI produce confident, wrong answers, and they don't want their name on that work.
These are the same losses that drive resistance to any significant change, which we cover in our guide to resistance to change. They are rational from where people sit, and they respond to the same things: honesty about what is changing, a say in how it happens, and support while people learn.
We saw the competence worry at a client whose staff were moving from paper to a new software system. Most had built their working lives without computers, and experienced people had to become beginners again. One-on-one time, a safe place to make mistakes and lifting work from the people who needed longer made the difference. We tell that story in the technology company that ran on paper.
Do leaders and staff see AI the same way?
No, and the gap is wide. SHRM's 2026 survey of 5,875 US workers found that 80% of directors and above trust senior leaders when they talk about AI. Among individual contributors, the figure drops to 47%, and another 38% are unsure. The same survey found that 74% of directors and above were told about possible AI changes before they happened, compared with only 33% of individual contributors (SHRM 2026).
The people furthest from the decisions are usually closest to the work AI will change. They hear about it last, trust the story least, and are then asked to make it work. Leaders who are confident about AI can mistake their own enthusiasm for the organisation's. The fix is mostly in the timing: tell people earlier, and bring some of them into the decisions.
What happens when the official tools lag behind?
People use their own. Microsoft and LinkedIn's 2024 survey of 31,000 people across 31 countries found that 75% of knowledge workers were already using AI at work, and 78% of those users were bringing their own AI tools. In the same survey, 60% of leaders worried that their company lacked a vision and plan to implement AI (Microsoft and LinkedIn 2024).
Australian workplaces follow the same pattern. In the Melbourne and KPMG study, 49% of Australian employees said they use AI regularly in their work. Almost half of those using it (48%) admitted to doing so in ways that contravene company policy, and 60% said they had hidden their use of AI at work (Gillespie et al. 2025).
Hidden use is a leadership signal. It usually means people find the tools useful, the approved options are missing or hard to use, and staff don't feel safe being open about it. Banning the tools tends to push the use further out of sight. A better response is to make the approved tools easier to reach than the unapproved ones, set clear and sensible rules about data, and make it safe for people to say what they are using and why. People raise problems and admit mistakes when they believe they won't be punished for it (Edmondson 1999), and that applies to how they use AI as much as anything else.
Why fix the process before adding AI?
AI speeds up whatever process it sits on. If the process is slow, unclear or full of workarounds, AI tends to produce the same problems faster. We see a common pattern: a business with error-prone processes decides the answer is more AI, buys the tools, plugs them into the workflow, and finds the errors arriving sooner.
The organisations that get value map their processes first, fix what is broken, and then add AI where it clearly helps. Michael Hammer made the same argument about computers more than thirty years ago: don't automate, obliterate (Hammer 1990). His point holds. Technology applied to a bad process gives you a faster bad process.
Mapping the process also does useful change work. The people who do the job have to explain how it really works, which surfaces the workarounds nobody wrote down. It gives them a direct part in deciding where AI fits, which makes them far more likely to use it.
Where does AI help, and where does it get in the way?
The evidence suggests AI helps most where tasks suit it, and can make work worse where they don't. Telling the two apart takes the judgement of the people who do the work.
In a study of more than 5,000 customer support agents, access to an AI assistant increased the number of issues resolved per hour by 15% on average, with the biggest gains in speed and quality among less experienced and lower-skilled workers (Brynjolfsson, Li and Raymond 2025). In a field experiment with management consultants, those using AI on tasks within its capabilities completed 12.2% more tasks and finished 25.1% faster, with higher quality work. On a task outside those capabilities, consultants using AI were less likely to reach the correct answer than those working without it (Dell'Acqua et al. 2023). The researchers called this boundary a "jagged technological frontier", because it is uneven and hard to see in advance.
For leaders, the lesson is practical. Nobody can tell from the boardroom which tasks sit inside the frontier. The people doing the work can find out, if they are given time to experiment and permission to report what doesn't work.
How do leaders lead AI adoption well?
Leaders who bring people with them treat AI as a change in how people work, and lead it accordingly. In practice:
- Start with the problem. Name the work you want to improve before choosing the tool. "We want to cut quoting time" gives people something to aim at. "We are rolling out AI" doesn't.
- Say what you know and what you don't. If roles will change, say so early. If you don't know yet, say that, and say when you will.
- Tell people before you decide. Bring frontline staff into the decisions about where AI fits, rather than announcing the result.
- Map and fix the process first. Then add AI where the people doing the work agree it helps.
- Make the approved tools the easy option. Fast access, clear rules on data, and no penalty for admitting what people were already using.
- Train on real work. People learn fastest when they use a tool on their own tasks straight away, with someone to ask.
- Use it visibly yourself. Staff watch whether their leaders use the tools, and how honestly they talk about the results.
- Share the gains fairly. Be clear about what happens to the time AI saves. If it only ever means more work, people will notice.
- Watch the load. AI often arrives on top of several other changes. Our guide to change fatigue covers how to count and manage the total.
Is AI adoption an adaptive challenge?
Mostly, yes. Choosing tools, setting up access and securing data are technical problems that experts can solve. Getting people to rethink their roles, trust new ways of working and let go of expertise built over years is adaptive work, and only the people involved can do it (Heifetz, Grashow and Linsky 2009).
Leaders who treat AI purely as a technology rollout tend to get expensive tools that few people use well. Leaders who treat it as adaptive work spend more time listening, experimenting and pacing the change, and they usually get further. Our guide to adaptive leadership covers how to tell the two kinds of problem apart and how to lead the second kind.
If your organisation is introducing AI and wants its people to come with it, talk to us about our change management work. Our leadership workshops help leaders build the skills to lead change of this kind, and our guide to change leadership covers the wider picture.