This “AI adoption” talk is starting to sound like a comforting lie. Not because AI isn’t useful. It is. The lie is the idea that the main problem is people don’t have the right skills yet. Most workplaces don’t have a skills problem. They have a system problem. And throwing AI tools into a broken system just lets the system break faster.
The idea I saw framed it cleanly: the real question isn’t whether employees can learn AI. It’s whether the organization can actually absorb it. That word—absorb—matters. Adoption is buying software and running a training. Absorption is changing how work flows so the new power doesn’t get trapped in old habits.
Here’s the uncomfortable part: a lot of employees are already ready. They’re curious. They’re using tools on their own. They’re finding ways to draft faster, research faster, summarize faster, and get to the “thinking” part sooner. But then they hit the wall: approval chains, outdated rules, managers who measure effort instead of outcomes, and processes built for a world where every sentence had to be typed from scratch.
So you end up in this weird place where people have a capability upgrade, but the organization has no place to put it.
The promise is real, though. When AI takes the first rough pass off your plate, you can spend more time on judgment: what matters, what to prioritize, what to cut, what the customer actually needs, what the risk is. That’s higher-value work. That’s the work most companies claim they want people doing.
But companies also love control. And control hates speed.
Imagine you’re a project manager. AI helps you turn messy meeting notes into a clean plan in ten minutes. Great. Except the company still requires a weekly status template, three meetings to “align,” and a manager who only trusts work that looks painful. Your output gets better, but the system punishes you for being faster. So what do you do? You slow down. Or you hide the AI use. Or you burn out from doing “real work” plus the theater that proves you worked.
Or say you’re in customer support. AI could draft responses, find policy details, and help you de-escalate. But your team is graded on handle time and script compliance, not on actual resolution. So the tool becomes a way to move tickets, not solve problems. The customer loses. The company loses later when churn shows up. The only “winner” is the dashboard that looks clean.
The post’s point about “systemic redesign” is where the fight is going to be. Because redesign means leaders have to give up some comfortable myths. Like: productivity is hours at a desk. Like: mistakes are always personal failure instead of process failure. Like: if we just train everyone harder, the org will magically change.
No. If you want AI to unlock potential, you have to change what you reward, how you review, and how decisions get made.
And that’s exactly why most companies won’t do it.
They’ll buy the tools. They’ll run workshops. They’ll celebrate “innovation.” But they won’t touch the messy parts: who has permission to decide, how risk is handled, whether people can experiment without getting smacked for one bad outcome. Absorption forces you to confront your own management style. Adoption doesn’t.
There’s also a real power angle here. If AI helps individuals produce work that used to take a team, the organization has a choice. It can use that gain to create more value—better products, better service, more creative work. Or it can use it to squeeze headcount and call it “efficiency.” One path builds loyalty and momentum. The other path teaches everyone the same lesson: never show your full capacity, because it will be used against you.
Some leaders will argue that moving slowly is responsible. That you can’t just let everyone run wild with new tools. Fair. There are risks: errors, privacy, bad outputs, messy accountability. But “move slowly” often becomes a cover for “we don’t want to change how we run things.” Meanwhile, employees don’t stop. They just do it quietly. And that’s actually worse, because now you have shadow habits with no shared standards.
If you’re serious about this, the culture part isn’t a poster on a wall. It’s practical. Can a person try a new workflow without needing five approvals? Can teams share what works without it turning into a compliance nightmare? Can managers judge work based on results and reasoning instead of busywork signals? If not, the tool sits on top like a glossy layer, and nothing underneath changes.
The organizations that will outperform aren’t the ones with the flashiest AI announcements. They’ll be the ones that redesign the boring stuff: how work moves, how learning happens, how people get unblocked, how good ideas spread, how risk is managed without choking everything.
And here’s what I’m not totally sure about: whether leaders will choose redesign willingly, or only after competitors force their hand and the old way starts losing money in a visible, painful way.
So what should matter more right now—giving every employee better AI tools as fast as possible, or slowing down to rebuild the systems those tools will run into?