The scene is becoming increasingly familiar. The company bought the AI licenses about a year ago. There was a launch event, a lunchtime webinar, and a few people were named “AI champions.”
Now someone puts the usage dashboard on the screen. A small group uses the tools every day, and most people have barely opened them. After a short silence, someone asks the question I’ve learned to expect: “Did we buy the wrong tool?”
They almost never did. They bought a perfectly good tool and expected people to work out the rest on their own.
When McKinsey asked leaders how far along they were, only 1% said AI was fully built into how their company works and was producing real business results. The technology may be advancing quickly, but the organizations adopting it are struggling to keep up.
We’ve done this before
When electricity reached factories in the late 1880s, people expected productivity to jump. It took decades. Factory owners replaced the steam engine with an electric motor and kept everything else as it was, including the layout, the workflow and the way people were organized.
Many companies are doing the same with AI. They take a process that was designed fifteen years ago, add a chatbot to it, and call that transformation. The work gets done a bit faster, but it still goes through the same approvals and the same bottlenecks as before.
The step everyone skips
Most organizations assume that once people start using AI and the company has invested in the technology, transformation will follow.
There’s a stage in between that hardly anyone plans for, and I call it AI readiness. Leaders need to agree on why they’re doing this at all. Managers need some idea of how to run a team where half the work now goes through an AI tool. Employees need to know when they can trust what the tool gives them and when to check it. And someone needs to reconsider the processes, expectations and measures of performance that were designed long before AI arrived.
None of that comes with the license.
One CHRO told me the technology was never the hard part. Getting her own leadership team to agree on what AI meant for the business took far longer.
Klarna found this out in public. In early 2024, it announced that its AI assistant was doing the work of 700 customer service agents. About a year later, its CEO admitted the cost-cutting had gone too far, and the company started hiring people again so customers could always reach a human.
Your people are further ahead, and more worried, than you think
In the same McKinsey research, executives estimated that 4% of their employees used generative AI for at least 30% of their daily work. When the employees were asked, the figure was 13%. The gap suggests that leaders may have far less visibility into how AI is actually being used than they think.
I have seen a similar pattern in organizations I’ve worked with. A few people subtly use AI every day, while colleagues doing much the same work barely touch it. Those who have become proficient rarely talk about it, while others are reluctant to admit they’re struggling to keep up. The result is a growing divide in AI confidence and capability, often within the same team.
Microsoft’s 2026 Work Trend Index found that 65% of AI users worry about falling behind if they don’t adapt quickly, but only 13% feel rewarded for reinventing the way they work. Most people are willing to change, but the way they’re measured and rewarded hasn’t caught up.
Five questions for your next leadership meeting
Before you approve the next AI tool, answer these honestly.
1. Are we telling the same story about AI?
If you asked every member of your leadership team to explain the company’s AI strategy, would you get the same answer? If not, managers are left to interpret it for themselves, and employees are getting mixed messages.
2. What problem are we trying to solve?
Asking “Where can we use AI?” starts in the wrong place. Start with a business problem that matters, then ask whether AI can help with it.
3. Would we design this process the same way today?
If the answer is no, rethink the process rather than simply automating what already exists.
4. Do our people know when to question the output?
Learning to write prompts doesn’t take long. Knowing when to trust AI and when to rely on your own judgement takes much more practice, and it matters more.
5. What would we say if someone asked whether their job was safe?
You may not be able to promise job security. You can promise to tell people early, to give them time to learn during working hours, and to be honest about what you don’t know yet.
A word to the people who signed the AI budget
When the usage figures disappoint, the easy thing is to blame employees. They’re too slow, too set in their ways, too nervous about change. I’ve heard all of it, and in most cases it’s wrong.
People are unlikely to embrace AI when nobody has explained why it’s there, when their own manager isn’t using it, and when the only time set aside to learn was a webinar over lunch. The problem usually sits higher up the organization, and so does the fix.
So before you approve the next AI investment, ask a harder question than whether it’s the right tool. Ask whether you’ve done the work to get your people ready for it. And whether you’re prepared to change the way the organization works to make that investment worthwhile.
If the honest answer is no, keep your money. You’ll only be looking at the same dashboard again next year.
