How to AI — What Really Works
There's a whole industry telling organisations how to AI right now. Buy the platform, appoint the champions, cascade the comms, count the licences. I've read a lot of the playbooks. Here's the bit that keeps getting missed: the technology is the easy part.
Successful AI adoption is rarely a technology challenge. It's a confidence challenge. People need to know what these tools can actually do, where the risks sit and how to use them responsibly — and no amount of licence procurement gets you that. You can roll out the shiniest assistant on the market and still find, three months later, that half your organisation is quietly not using it and the other half is using it in ways that would make your risk team go pale.
Show, don't announce
What really works is showing people, in context. Not a white paper — a Wednesday-afternoon five-minute time saver. When we ran informal drop-in sessions, colleagues demonstrated how they turn meeting transcripts into plans, redraft an email in a friendlier tone from a handful of bullet points, and summarise dense research before the kettle's boiled. Attendance spiked, questions flowed and confidence visibly grew. When technology is shown in context, barriers turn into curiosity.
Visible sponsorship matters more than sponsors tend to think, and not in the way they'd like. The most useful thing a senior leader can do is fumble through a live demo. The polished keynote about transformation changes nothing; the moment a leader shares their own wins and wobbles, everyone else gains permission to experiment. That's how organisations learn, quickly and safely.
Governance should make people braver
This is where my day job comes in, and where governance gets an unfair reputation. Done badly, it's a queue. Done well, it's the reason people feel safe to try things: clear boundaries, plain-language guidance, and someone accountable for every system. Treat AI literacy as a control, not a nice-to-have. When people understand what they're accountable for, fewer things go wrong upstream — long before any policy has to catch them. No owner, no oversight.
None of this is anti-technology; the opportunity is why I work in this field. But AI used well is a sharp, critical peer, not a lazy substitute for thinking. So keep the humans central — their confidence, their judgement, their accountability. Get that right, and the rest of the transformation has something solid to stand on.