A while back I wrote about the math that keeps me up at night — how organizations will approve a six-figure AI platform without blinking, then hesitate over a fraction of that for the people who actually have to use it. The argument was simple: you have to fund the people more than the platform, or the platform never delivers what you bought it for.

That was the theory. Since then, we’ve run the real thing.

Over the past few months, UnconstrainED delivered hands-on AI training to four companies across four industries: a boutique hotel group, a diamond and jewellery house, a cross-border finance and accounting firm, and a global elastics manufacturer. Different countries, different maturity levels, different problems. Cross-functional rooms every time — sales, operations, finance, HR, IT, manufacturing, marketing, reservations, admin, all in the same session.

I want to tell you what we learned. Not the highlight reel — the thing that actually surprised us.

The demand is not the problem

Here’s what I expected walking in: some resistance. A few skeptics with their arms crossed. The usual “why do we even need this” energy.

That’s not what we found. Not even close.

What we found was hunger. People leaning in, staying past time, asking for the next thing before we’d finished the current one. Across every cohort the pattern repeated: enthusiasm consistently ran ahead of the infrastructure around it.

At the elastics manufacturer, the single biggest thing standing between people and progress wasn’t motivation. It was corporate IT access. People wanted to build, wanted the tools in their hands right now — and couldn’t get them unlocked fast enough. Sit with that for a second. The bottleneck wasn’t reluctance. It was that the organization couldn’t move as fast as its own people wanted to.

The research says the same thing at scale. The World Economic Forum’s Future of Jobs 2025 projects that 59% of the global workforce will need reskilling or upskilling by 2030 — and 11% won’t get it. Employers now name skill gaps the single biggest barrier to transformation. The appetite is there. It’s the follow-through that isn’t.

If you’re a leader waiting for your people to “be ready” for AI: they’re already ahead of you.

Then why does so much training fail?

Because most of it is generic. And people can smell generic.

A recent Docebo report put numbers on something I’ve watched play out in room after room: 85% of employees say the AI training they receive doesn’t actually help them use AI in their role. 78% say it happens in tools disconnected from where they really work. Nearly 80% say it isn’t personalized to them at all.

That’s the whole problem in three statistics. Not a shortage of training — a shortage of training that connects to the actual job. Watch a slick demo on a made-up dataset and you learn nothing you can use on Monday. It’s tool tourism. You saw the sights; you still can’t navigate the city.

That gap is the thing we built our entire model to close.

What actually made it land

When we debriefed across all four clients — more than 110 feedback responses in total — the same three things came back, again and again.

It was grounded in their real work. At the finance firm, participants didn’t watch us demo on a sample spreadsheet. They arrived with their live problems — invoice tracking, variance analysis, reconciliation workflows — and those became the actual build briefs for the session. They walked out with something they’d use the next morning, because it was the next morning’s work. This isn’t just a preference; it’s how people learn. The largest meta-analysis of active learning found students in traditional lectures were 1.5 times more likely to fail than peers who learned by doing. You don’t build capability by watching. You build it by building.

We taught a framework, not a tool. Tools change every month. Judgment doesn’t. At the elastics manufacturer, 77% of participants named our 4Ps prompting framework — a simple way to structure any request to any AI — as the thing that landed hardest. Same story at the hotel group. Give someone a framework and they can teach themselves the next tool. Give them a tool and they’re back in your inbox the moment it updates.

It built, session over session. This wasn’t a one-off workshop that evaporates by Friday. At the manufacturer, by their second session one cohort had caught up to where a more advanced cohort had started — and in that advanced group, over 70% named AI agents and projects, the genuinely harder material, as the most useful part. That’s not people rating a session politely. That’s people moving.

The receipts

I’m wary of case-study numbers, because anyone can cherry-pick a happy quote. So here’s the aggregate, plainly:

• The diamond and jewellery house: 9.1 out of 10 for a half-day session.
• The manufacturer’s first cohort: 4.4 out of 5, with 75% giving top marks.
• Recommendations: every cohort but one said they’d recommend it — unanimously. The single exception still came in at 94%.

Four industries. Cross-functional rooms. Over a hundred responses. The consistency is the point — this wasn’t one charismatic session that happened to click. It was the same result in a hotel, a diamond house, a finance firm, and a factory.

What this actually tells you

The story the market tells itself is that AI adoption is hard because people resist it. Our field data says the opposite. The people are ready. In a lot of cases they’re impatient. What’s holding organizations back isn’t their workforce’s willingness — it’s whether leadership builds the capability and clears the path. Yes, including getting IT to unlock the tools before the enthusiasm cools.

The tools were always going to be the easy part. I said that last time, and I’ll say it again, because four industries just proved it to me in a row: the people were always going to be the work — and they’re far more willing to do it than you think.

One question to sit with this week

If your people are already using AI — and the odds are they are — are you handing them training built on their real work, or a demo on someone else’s?

We built a free Workflow Audit to help you scope exactly that: where AI actually fits your team’s day, using their tools and their tasks. Use it on your own, or bring us in to run the room.

If this is the conversation you’re having in your own org, I’d genuinely like to hear how it’s going. alex@unconstrained.work.

— Alex