Why Firing Your Way to AI Adoption Backfires (And What Actually Works)

A CEO recently made headlines for laying off nearly 80% of his staff because they refused to adopt AI fast enough, and two years later, by his own account, it worked out for the business. It's a striking story, and almost useless as a playbook for a Malaysian SME weighing how to actually drive AI adoption.

The trouble with the mandate-and-purge approach is that it's survivor bias dressed up as strategy: for every leader who fires their way to adoption and it happens to work, there are far more who lose their institutional knowledge, their best people, and their culture, and still end up with an organisation that uses AI badly, just with fewer people left to notice.

The counter-argument that actually holds up

A more useful piece of business writing I came across recently makes the opposite case, and it holds up: stop asking employees to "adopt AI" as if that's a single, universal instruction. It isn't a task anyone can actually execute. It's a slogan. Real adoption happens when a specific person can see how AI helps with a specific, annoying part of their specific job, not when leadership announces a company-wide mandate and hopes enthusiasm follows.

This matches what I see consistently in Malaysian SMEs. The businesses where AI adoption sticks aren't the ones with the most aggressive top-down directive. They're the ones where someone in operations found a genuinely useful application, cutting an hour off weekly reporting, catching errors a tired reviewer would miss, and other people noticed and asked how.

Why mandates fail even when compliance is high

There's a real difference between employees using a tool because they're told to and employees using a tool because it makes their day genuinely easier. The first produces box-ticking, a login used once a month to satisfy a metric. The second produces the compounding, habitual use that actually changes how work gets done. A mandate can force the first. It cannot force the second, and the second is the only one that pays off.

This is also where the layoff approach quietly fails on its own terms: if the goal is genuine capability rather than headcount optics, removing the people who were slow to adopt doesn't solve the underlying problem, it just removes the evidence of it. Those employees were often responding rationally to tools that hadn't yet been made relevant to their actual work.

What this looks like in practice for a Malaysian SME

1. Find the one task, not the whole department. A single, visibly annoying task solved well beats a company-wide rollout announcement every time.
2. Let the early adopters be the messengers, because colleagues trust colleagues more than they trust a memo from leadership.
3. Budget for the awkward middle: the period where some staff are faster with AI than others. That gap closes with coaching, not with headcount reduction.

What This Means for You

The layoff story gets attention because it's dramatic. Stopping the mandates and solving specific problems instead gets results because it's true. If you're leading AI adoption in a Malaysian SME, that's the playbook worth following, even though it makes for a much less exciting headline.