You rolled AI out across the company, and it isn’t being used the way you expected. So the question now is what to do next.
The standard explanation goes like this. Your people are afraid. If they get caught using it, their rating drops; if they deliver too much, their work gets taken away. So they either use it in secret or don’t touch it at all.
The 2024 Work Trend Index from Microsoft and LinkedIn (31,000 knowledge workers across 31 markets) does back up part of that story. Among people who already use AI at work, 52% are reluctant to admit to using it for their most important tasks, and 53% worry that using it on important work makes them look replaceable. And yet 78% of those AI users bring their own AI tools in, outside whatever the company handed them. At least the reluctance to say so out loud is real. The stage before this one — where leadership keeps chasing tools and the rollout itself never happens — is a separate piece. Today is about what happens after you’ve handed it out.
Nobody here is hiding anything
I tried to apply that story to our team in Vietnam and it didn’t fit. I asked our people whether there was an atmosphere where using AI would hurt their rating. What came back was, “No, nobody has given off that feeling so far.” I’ve never caught anyone hiding it. Of course, they may just be saying that because the boss is the one asking, and anyone hiding it is invisible to me by definition. But AI usage gets shared openly in our chat every day, and I struggle to find a reason anyone would hide. If anything it runs the other way: I keep telling the team, let’s become the number one company in Vietnam at actually running AI, and that we lose that race unless we free up time with AI. That’s a flag to charge behind, not a surveillance camera.
And still, the team split cleanly into people who use it and people who don’t. No fear in the air, and the split happened anyway. That’s what this piece is about.
The data behind the fear story doesn’t put fear first
There are actually two different things packed into the word “fear”: anxiety about how your usage looks, and simply not having time to touch the thing. Leave them mixed and the conversation drifts, so let me re-read the received wisdom’s own data with the two pulled apart. Slack’s Fall 2024 Workforce Index (17,372 desk workers across 15 countries) found that nearly half — 48% — would be uncomfortable telling their manager they had used AI on a work task. Among that 48%, the reasons cited most often were: using AI feels like cheating, 47%; fear of being seen as less competent, 46%; fear of being seen as lazy, 46%.
Job security isn’t on that list at all — though the survey never offered it as an option, so this is an absence of measurement, not a low ranking. What is on it is how you look to the people around you. But that kind of fear is a wall in front of disclosing. The wall in front of touching it at all is probably a different wall. In a place like ours where nobody hides, the people who don’t use it still don’t use it. Which means a design that starts by promising job security may, at least on this data, be aimed at the wrong wall.
So what is the real wall? The same survey has another set of numbers. 61% have spent under five hours total learning AI. 30% have never had any training. Only 7% consider themselves expert AI users.
What those three describe is something far more mundane: almost no time has been put into touching it. Whether that’s because time couldn’t be found or because it simply wasn’t a priority, these numbers can’t tell us.
Divide one, the time divide: can you block a real chunk of time?
This is what was actually happening at our company. In the end, the people who use it use it, and the people who don’t, don’t. Whoever can block a real chunk of time uses AI to move up a level, then takes more time and moves up again. Whoever can’t block that time doesn’t get any lighter from touching AI, so eventually they stop.
People who genuinely enjoy playing with AI will carve out the time themselves and walk into that loop with no help. People who are more passive have no trigger to enter it unless the company blocks the time for them. So what separated the two groups wasn’t motivation and it wasn’t fear. It was whether the time to enter the loop had been secured.
Divide two, the method divide: making your own work easier, versus handing that ease to others
Clearing the time divide and entering the loop doesn’t mean everyone rises to the same height, though. There’s another wall past it. We’re a software company, so our engineers are, by nature, people who want their work to be easier. Which means “use AI to make my own tasks lighter” happens on its own, without anyone pushing.
The layer above that — the method layer — is where it gets hard. Make it easier, then turn that ease into a method other people can run. Concretely: take the lighter way you found and put it in a form you can hand to someone else. So far, the only people doing that are the ones who can put their own method into words.
I don’t think that’s a gap in technical skill so much as a gap in whether you have a method of your own. How you break your daily work apart, how you run it. People who hold that personal golden framework can share it with AI. People who’ve never put their own way of working into words have very little to hand over, so the collaboration spins in place. This runs straight into the piece about “just tell me what to do” getting overtaken by AI.
What leadership builds first is the system that blocks the time
You might read all this and ask whether that means the evaluation and reward design is unnecessary. I wouldn’t say unnecessary. I’d say the order is backwards. The mechanism that blocks time comes first, and the reward design comes after.
This one is about me. If you’re going to tell your staff to win with systems, put a system into your own work first. I had to point that at myself recently. I’d been telling the team to systematize everything, while my own sales activity had no system at all — and when I went back through the chat archive, there was a stretch of about seven weeks with zero sales activity. The person outside the loop was me.
What fixed it was a shape for the time. Block “sales, 30 minutes” on the calendar every Wednesday at 14:00. Put “last sales activity: X days ago” into my morning briefing, with a caution mark at 7 days and a warning at 14. That’s all. Build the mechanism that secures a real block of time automatically, before relying on anyone’s motivation. The contents of that block are sales, not AI, but the shape — fix the time first, automate the behavior — is the same as the time divide. The staff-facing version of this is still at the stage where I’m testing it on myself. Once I can see it working, the same shape goes to the team. After that comes the nudge toward the method divide: say “try writing down how you work right now” and watch whether the number of people who can hand their method over goes up.
If you would rather bring that “secure the time first” part in from outside, that is the service we run as AI Training for Teams, written up on a separate page. It’s the option for when you can’t get the blocks onto the calendar on your own.
On top of that, I think the flag should point forward rather than backward. “Don’t worry, using AI won’t cost you your job” is a defensive message; the charging flag I described at the top does more to motivate people to protect a block of time. And there are at least numbers behind the idea that leadership raising a flag at all matters. In Gallup’s Manager Support Drives Employee AI Adoption (2025, 19,043 US workers), inside organizations that are already investing in AI, employees who strongly agree their manager actively supports their team’s use of AI are 2.1 times as likely to use AI a few times a week or more. That is an association, not proof of cause. But a gap that size suggests a manager who visibly backs the team’s use of AI moves the needle more than most things a leader can do from a distance. I think raising the flag is the leader’s job.
There are still places where the received wisdom is right
Everything above argues “it’s time, not fear,” but that may only hold because we’re an IT company in an industry where engineers are short-staffed and in demand everywhere. In industries where the jobs themselves are genuinely at risk, the standard story — the fear that AI will replace me — is straightforwardly right in plenty of cases.
What happened at Commonwealth Bank of Australia in July and August 2025 is the example (reported by ACS Information Age). In July the bank announced that 45 customer service roles would be cut and replaced by an AI-powered voice bot, then reversed the decision in August, saying its initial assessment “did not adequately consider all relevant business considerations.” It apologized to the employees concerned. The finance sector union says call volumes to those workers were rising at the time and managers were offering overtime to cope. That said, what backs the fear story isn’t the eventual reversal so much as the shock the announcement itself sent through the floor. Talk about AI in the same breath as headcount cuts and people brace. Of course they do.
So the first thing to establish is which side your own company is on. If the job insecurity is real in your industry, say the job-security part out loud first. If you’re short-staffed and AI making work lighter would be welcomed, building the time-blocking system first works better. We’re doing the latter, and I don’t assume that generalizes to every company. One step further back: if you’re still stuck at “I don’t even know what we’re allowed to put into AI,” then the rules for sorting your data come before any time blocking.
FAQ
If employees don’t use AI, does that mean they aren’t motivated?
Not necessarily. What we can see at our company is a gap in whether someone can block a real chunk of time. Slack’s survey found 61% had spent under five hours total learning AI and 30% had never had any training. In many cases the time to touch it was never invested in the first place, well before motivation enters the picture — and those numbers can’t tell us whether that’s a shortage of time or a question of priority.
What does a time-blocking system actually look like?
Fixing a day and a time and putting it straight on the calendar is the fastest version. I fixed my own sales activity at “every Wednesday, 14:00, 30 minutes,” and set my morning screen to warn me when the activity had been idle for 7 or 14 days. The point is to make the time appear automatically instead of leaning on motivation.
How do we grow the method layer — the people who can hand their way of working to someone else?
Start by looking at whether each person can put their own way of working into words. Someone with no personal method — no golden framework of their own — has very little to hand to AI, and systematizing stalls. Block the time and get them touching it first, then nudge with “try writing down how you work right now.” That order feels natural to me.
Job insecurity is real in our industry. Is time blocking still first?
No — the order changes there. In industries where jobs genuinely are at risk, there are real situations where saying the job-security part out loud has to come first. Working out which side you’re on is the first step. One rough test: whether your people are in demand on the job market.
Let’s work out which side your company is on
Give me 30 minutes and we’ll sort out whether your company is a “block the time first” case or a “say the job-security part first” case — and what the first block on the calendar looks like once you know. Treat it as a sounding board session; no preparation needed.
Shogo Harada原田 祥吾
CEO · Linnoedge Inc. · LinkedIn↗
Operating IT offshore development and overseas expansion support businesses across two bases: Tokyo and Vietnam. A leader who believes in “Systems over Spirit,” structuring cross-border businesses that often tend to be opaque. Committed to providing “reproducible quality” to organizations and clients rather than relying solely on individual skills.