The wrong question about AI is getting more expensive to ask.

I wrote recently that AI won’t replace you, it will expose you, but, there’s a second shift underneath that, something I’ve been seeing in day to day work.

I see it in product reviews where the draft strategies are generated in seconds. In leadership meetings where analysis that used to take days arrives fully structured in minutes. In design sessions where options are no longer scarce.

You describe an outcome in plain English and receive something coherent back. It’s structured, polished and it looks great. That changes the economics of work. Most work falls into three categories - insight, throughput and judgment.

Insight is reframing the problem before anyone solves it, connecting incomplete information into a direction. Throughput is sustained volume, processing and delivering at scale without fatigue.

Both of those are being compressed by AI. You can now generate ten plausible strategies before most teams have finished defending their first one. You can process information at a scale that used to require headcount. But judgment, although influenced by AI has not moved.

Judgment is accountability under pressure. It’s backing a direction when the data is incomplete. It’s narrowing scope when everyone wants expansion. It’s carrying the consequence when the information is incomplete.

AI does not do that. Not because it isn’t smart enough, but because it isn’t accountable and it cannot absorb consequence. Intelligence can be replicated, but accountability cannot be delegated.

In a recent conversation with a CEO, we were mapping where AI genuinely improves what we’re building versus where it’s cosmetic. He used AI a lot and I asked him what AI actually gives him. He said, context. And that’s the right use. Not answers, context.

I had already been writing about this shift, but hearing it used instinctively, without theory attached, confirmed everything. The compression of insight and throughput by AI doesn’t eliminate expertise. It shines a light on your value.

If your value was assembling, formatting, producing, you’ll feel pressure. If your value is framing, prioritising and deciding, you become more valuable and you gain leverage.

The real divide now isn’t between people who use AI and people who don’t, it’s between people who use it to accelerate output and people who use it to sharpen judgment. One scales activity, the other scales consequence.

Look at your actual work. How much is insight, how much is throughput and how much is judgment? Because as AI gets cheaper, your value is no longer assumed, it is revealed.