AI isn't the risk. Forgetting people is.

The technology question is the easy one

Every AI strategy deck looks the same right now. A slide on the tools. A slide on the roadmap. A slide on governance and risk. Somewhere near the end, a single bullet point that says “upskilling”  as if it’s a footnote rather than the whole point.

It isn’t a footnote. It’s the strategy.

Choosing which AI tools to roll out, how to license them, and where to plug them into existing workflows is, relatively speaking, the simple part. It’s a procurement problem with a project plan attached.

The harder question, the one most organisations are quietly avoiding is whether their people actually have the confidence, judgement and skills to use any of it well. Not “can they log in and try it.” Whether they know when to trust an output and when to challenge it. Whether they can spot a plausible-sounding answer that’s quietly wrong. Whether they’ve got the judgement to know where AI stops and their own expertise has to start.

That’s not a licensing problem. That’s a capability problem. And it doesn’t show up on a rollout timeline.

Skills that matter more, not less

There’s a common assumption that as AI takes on more routine, repetitive work, the human skills around it become less necessary. It’s the opposite.

As AI absorbs the routine, the things left to people are the things AI can’t do:

• Adapting when the plan changes
• Communicating clearly when a decision is contested
• Making a judgement call with incomplete information

Strip away the routine work and what’s left is more human, not less — which means any skills gap gets more exposed, not less.

Managers feel this first. They’re the ones being asked to lead teams through a change they may not fully understand themselves, to answer questions about job security they can’t fully answer, and to keep performance up while the ground is genuinely shifting under people’s feet. Leading through that uncertainty is its own skill — it doesn’t appear automatically because someone got a manager title five years ago.

A tools rollout isn’t a capability strategy

None of this is an argument against AI adoption. It’s an argument against treating adoption as if the technology is the whole job.

An organisation can have the best AI tools on the market and still lose the value if the people using them lack the confidence to use them well, the judgement to know when not to, and the communication skills to bring the rest of the organisation with them. The risk was never really the AI. It’s building a strategy around the tool and forgetting the people who have to make it work.

Where to start

If this is landing because it’s already on your agenda, you’re not alone – it’s the conversation we’re having with L&D leaders right now. The organisations getting ahead of it aren’t the ones with the flashiest tools. They’re the ones investing as seriously in adaptability, communication and decision-making as they are in the technology itself.

If you’re thinking about the people and capability side of your AI strategy, we’d like to talk.

Get in touch with Jam Pan to talk about the people side of your AI strategy.