We bought the tools and nobody uses them. How do we drive adoption?

We bought the licences six months ago and almost nobody logs in, and you do not know whether to push harder or quietly cancel. The invoice renews next month. The champions who lobbied for the rollout have gone quiet. Somewhere in a finance review, someone is going to ask what the return was, and you do not have a clean answer.

Here is the answer to the key question. You drive AI adoption by treating it as a change in how people work, not as a feature you have already bought. The licences are dormant because the tool was provided without an associated workflow, no coaching, and no answer to the quiet questions your people are carrying. The fix isn't more pressure on access and it is not a cancellation. It is sponsoring specific workflows, visibly, and walking the people through them. The technology was never the hard part.

Why did everyone agree the tools were brilliant and then stop using them?

Because access is not adoption, and the gap between the two is where most of the money quietly dies. Start with the broader pattern. Zylo, a firm that tracks software spending across more than 40 million licences, found in its 2025 SaaS Management Index (an annual study of how companies buy and use cloud software) that 52.7% of purchased licences sit idle in a given month, wasting an average of 21 million US dollars a year per organisation. Your dormant AI seats are not an anomaly. They are the median experience, captured on your own invoice.

Gartner saw the same shape coming. It predicted that at least 30% of generative AI projects will be abandoned after proof of concept by the end of 2025, on the back of unclear business value and rising costs. As their analyst Rita Sallam (a Gartner research vice-president who studies enterprise AI) put it, executives are impatient for returns while organisations struggle to prove and realise value. The diagnosis is consistent: licences activated with no onboarding or communication plan, a tool that asks for a behaviour change rather than a click, and unanswered trust questions that stall adoption more reliably than anything else. That is your situation, named.

So is the problem the technology, the people, or me?

It is the system around the tool, and that is good news, because the system is the part you can move. BCG (Boston Consulting Group, the management consultancy) has a number I return to often: the 10-20-70 rule. Roughly 10% of the effort and value of an AI transformation sits in the algorithms, 20% in the technology and data, and 70% in people and process: the roles, the workflows, the change management, the governance. You bought the 10%. The 70% is still sitting in the box.

This is what I mean when I say the bottleneck is no longer the technology. The licence works. The model is capable. What is missing is the human operating system around it, and most leaders are trying to install new software on broken hardware: a new tool dropped into an unchanged workflow, with no time, no coaching, and no signal from the top that this is how we work now.

Gartner put a sharp edge on it. Its 2025 Microsoft 365 and Copilot survey found that 40% of organisations were piloting Copilot and another 19% were planning to, yet only 5% of those that had finished a pilot moved to larger deployment in that year. Gartner named the cause plainly: enablement and change management, not the tool. Having the tools in hand does not lead people to find many uses for them, especially when the tools change weekly. The technology is in the building. Permission, direction and support are the parts still missing.

What does driving adoption actually look like, step by step?

The category of intervention here is deliberate enablement: a small set of sponsored workflows, real coaching, and a visible signal from leadership that this is the way we work now. BCG's AI at Work 2025, a survey of frontline employees and managers across many countries, makes the size of that lever clear. The share of frontline employees who feel positive about generative AI climbs from 15% to 55% when they sense strong leadership support, yet only about a quarter say they currently receive it. Regular use is sharply higher among the people who get at least five hours of training plus in-person coaching. Sentiment is not fixed. You set it.

And it compounds. The split between the firms pulling ahead and the firms stalling is not about who bought the best tool. It is about who led the adoption on purpose. Gartner's own survey makes the point: nearly everyone has access, almost no one has reached scale, and the difference is the human work in between. Here is the sequence I would run.

1. Choose two workflows, not the whole company. Pick two high-frequency tasks where a good output is obvious and the people doing them already feel the friction. Depth in two beats a thin layer across forty.
2. Sponsor them out loud. Name the workflows in a leadership forum, say why they matter, and commit to them in front of the team. This is the 15%-to-55% lever; spend it where everyone can see.
3. Coach the people, then keep coaching. Five hours of hands-on time with someone who sits beside them turns familiarity into a habit. A one-off webinar leaves it as a memory.
4. Answer the trust questions early. Where does my data go, will this be used to measure me, am I still valued. Named out loud, these clear the path; left unspoken, they quietly stall everything downstream.
5. Track the value, then expand. Measure time saved or quality gained in those two workflows, share the result, and let the next two be pulled by demand rather than pushed by mandate.

There is a quieter layer to this, for the leader ready to hear it. People adopt change from a settled nervous system, not a fearful one. The autonomic research is suggestive: higher resting heart rate variability (the natural beat-to-beat variation in your pulse, an index of how well the body self-regulates) tracks the prefrontal circuits that govern self-control and clear decisions, and even judges, in the well-known parole study by Danziger and colleagues, made far better calls when rested than when depleted. I mark the human-coherence reading of this as exploratory. The practical point is firm: a workforce led with calm clarity adopts. A workforce flooded with mandates and anxiety quietly logs off. The leaders who win the next decade are the ones who upgrade themselves first, and the room follows the state of the person leading it.

About the author

Thomas Green

British technology futurist, AI keynote speaker and advisor. Thirty years across enterprise technology and AI strategy, helping leaders navigate the future of work. The futurist who died.