Trish's Blog

Two numbers have stayed with me since I first read them, and I think every leader could do with paying more attention to them.

Ninety-nine percent of CEOs expect AI to reduce their headcount over the next two years. Only thirty-two percent believe their own organization is any good at combining human work with AI systems.

Read those side by side and a strange picture emerges. Almost every leader is confident enough in the technology to bet jobs on it. Almost none of them feel confident in the thing that determines whether that bet pays off, which is the join between the people and the machines.

That is not a small inconsistency. It is a hole in the middle of most AI strategies, and a lot of value is hiding inside it.

Here is the trap. When a technology arrives this quickly and this loudly, the instinct is to focus everything on acquisition. Which model, which vendor, which features, how fast can we deploy. These are answerable questions, and answering them feels like progress. You can put them on a slide. You can report them to a board. You can measure spend and adoption and feel that something is being done.

The harder question, the one only thirty-two percent feel they have answered, is what happens at the seam. How do people and AI work together so that the result is better than either alone? That question has no vendor. No one can sell you the answer, because it depends entirely on your people, your culture, and the way decisions are made in your particular rooms. So it gets quietly skipped, and organizations end up with excellent tools bolted onto unchanged habits.

I see the consequences regularly. A team adopts a powerful system and uses it to do the old thing slightly faster, rather than to do a better thing. The AI generates a confident answer and nobody feels equipped to challenge it, so a subtle error sails straight through. Or the technology takes over the routine work and, instead of freeing people to think, simply leaves them anxious about what they are now for.

In every one of those cases the problem is not the technology. It is the human side that nobody invested in. The tool was treated as the whole answer when it was only ever half of one.

This is where I get genuinely optimistic, because the hole in the middle is also the opportunity. The companies that close it will not do so by buying more software. They will do it by getting far more from the people they already employ. The same workforce, thinking better together with the machines, becomes the edge that no competitor can purchase.

What does closing it look like in practice? It starts with treating the human contribution as seriously as the technical one. That means teaching people not just how to use the tool, but when to trust it and when to push back. It means building the confidence to say “the system is wrong here, and here is why.” It means designing work so that the machine handles what it is good at and people are pointed at the parts that need judgement, context, and the awkward question a model would never think to ask.

It also means being honest about what AI cannot do. It cannot care about your customers. It cannot read the unspoken thing in a room. It cannot decide what matters. It is extraordinary at producing answers and useless at deciding which questions are worth asking. That deciding is human work, and it becomes more valuable, not less, as the answering gets automated.

The organizations betting their headcount on AI are making a wager about the future. Fair enough. But a wager you do not know how to win is just a risk. The thirty-two percent who feel ready are not the ones with the fanciest tools. They are the ones who have done the slower, less glamorous work of helping people and machines think well together.

If you are choosing where to put your next bit of effort, I would not put it into matching your competitor’s latest model. They will match yours right back, and round it goes. I would put it into the seam. Into the human side of the system, where almost everyone admits they are weak and almost no one is looking.

That is where the advantage is hiding. Not in the technology everyone has, but in the thing almost no one has figured out how to do with it.