You bought the AI. Did anyone’s work change?

Part of Measuring behaviour change.

The licences are live. The work looks the same.

The rollout went well. Licences were bought, logins were issued, an all-hands demo showed what the new assistant could do. A few months later, someone asks the obvious question in a leadership meeting: what has actually changed?

Often the honest answer is: not much. The reports are still built the same way. The same meetings produce the same follow-ups. A handful of enthusiasts use the tool every day; most people opened it a few times and drifted back to what they knew.

Companies are pouring money into AI tools. So why is everyone still working the same way?

What employees are telling us

On 9 October 2026, Cognizant published research on how employees relate to AI at work. Two findings stand out.

First, up to 82% of employees surveyed say their current AI tools don’t fit their needs [1].

Second, four in five say they want to use AI more than their job permits [1].

Read those together and the usual story falls apart. The problem isn’t that people are reluctant. Most of them want to use AI more, not less. What holds them back is the fit between the tool and the work, and how much room the job leaves for using it.

That is a useful correction for anyone who has treated slow adoption as a motivation problem to be solved with another launch email.

L&D has the same blind spot

It would be easy for L&D to read that as someone else’s problem. It isn’t.

In Donald Taylor and Eglė Vinauskaitė’s October report for OpenSesame, 68% of 531 L&D practitioners worldwide say their organisations now use AI in L&D [2]. That’s real adoption.

But the most common use is still making content [2].

Faster course production has its place. Yet it repeats the pattern above: the tool is in, and the thing it touches most is output, not how people work. More modules, produced more quickly, don’t on their own change what anyone does on Monday morning.

If the wider workforce is stuck between wanting to use AI and being allowed to, L&D is well placed to help. It can only do that if it looks past content and at behaviour.

Tools don’t change habits. Jobs do.

Behaviour shifts when a job expects the new way.

That sounds obvious, but it’s easy to forget during a rollout. A licence makes a new way of working possible. It doesn’t make it expected. Three things usually do:

• Practice. People need repeated, real chances to use the tool on their own work, not a demo on someone else’s example. One good attempt in a workshop rarely survives the first busy week.
• Feedback. Someone needs to tell them what worked, what didn’t, and what good looks like here. Without it, people either give up after a poor first result or settle into a shallow habit that never improves.
• A manager who notices. When a manager asks how the new approach went, builds it into how work is reviewed, and makes time for it, the message is clear: this is part of the job now. When nobody notices, the old way quietly wins.

None of those is a content problem. All of them are about how the job is set up.

What this means for your programmes

For L&D teams supporting an AI rollout, a few practical questions help:

• Where does the job actually expect the new way? Pick one team and one recurring task. Is using the tool part of how that task is done and reviewed, or is it optional?
• Where do people get to practise on real work? Swap one generic session for structured practice on the team’s own tasks.
• Who gives feedback, and when? Name the person and the moment. Practice without feedback is just repetition.
• What do managers see? Give managers a simple way to notice and talk about the new way of working in their regular conversations.
• What are you watching for? Look for changes in how the work gets done over a few months, not how many people logged in during launch week.

None of this needs a new platform. It needs someone to ask how quickly new tools turn into new habits, and what in the job helps or blocks that.

How fast do new tools become new habits?

So ask how fast your organisation turns new tools into new habits.

That is the question the Learning Velocity Index™ is built to look at. It puts Competence and Resilience in one score. It’s a self-assessment of how fast and sustainably an organisation builds capability, which is what an AI rollout ultimately depends on.

Like any self-assessment, it’s a first read rather than a final verdict, and a useful way to start the conversation. If you’d like a starting point, you can get your number, free.

Key takeaways

• Most employees in Cognizant’s research want to use AI more, not less. Weak adoption is often a fit-and-permission problem, not a motivation one.
• L&D’s own AI use is widespread but still centred on making content.
• Behaviour shifts when the job expects the new way: practice, feedback and a manager who notices.
• Judge an AI rollout by what changes in the work, not by licences issued.

Figures from the public sources linked. No affiliation with or endorsement by the organisations named.

General information on training design, not professional advice for your organisation.