AI is the current wave, not the work.

Why does every wave feel like the last chance?

We have been building technology for serious organisations since 1999. In that time the wave has been the web, then ecommerce, then mobile, then cloud, then big data. Each one arrived with the same three claims: this changes everything, the winners are moving now, and you are already behind.

Some of that was true each time. Most of it was noise. The organisations that did well were not the fastest adopters. They were the ones that made better decisions, managed risk intelligently, brought their people with them, and turned the technology into capability that lasted after the headlines moved on.

We say this with some history behind it. Our own AI story did not start with ChatGPT. It started in the 1990s, building the infrastructure behind AI systems at Inference, and continued through two decades of running machine-driven personalisation in production at national scale.

Where do the hype and the panic actually come from?

The same place: not knowing yet. Boards panic because they cannot tell what is real. Vendors inflate because uncertainty sells. Teams freeze because nobody has told them what they are allowed to do.

Once an organisation knows, the noise falls away. What is left is business as usual, with new benefits. A customer gets an answer the same day instead of next week. A team stops doing a task that never deserved a human. That is what success looks like: not a transformation announcement, just better ordinary work.

Is AI different this time?

It is big. It deserves serious attention and serious investment. But “different this time” is what every wave said, and the fundamentals have not moved: not every problem is an AI problem. Many of the problems we are asked to solve are traditional, fixable with simpler tools. More still are structural: people, teams and process. Recommending AI where AI is not the answer is how a partner loses the right to be trusted. If you want to test that in a room, these are the questions we would want you to ask us.

What should leaders actually do?

If the conversation is heading for a board, the shape of it matters as much as the content: what your board actually needs to know.

Three things, in order. First, get clear on where the genuine value sits for your organisation, and be honest about where it does not. Second, put governance in place that enables pace rather than blocking it, so your people can move without creating exposure. Third, measure adoption in behaviour, not in pilots launched. The test we use is simple: if we offered to take the new capability away, would your people fight to keep it?

The wave will pass. The vocabulary will change again. The capability you build, and the judgement you apply while building it, will still be working long after the hype has found a new subject.

Common questions

Is AI a bubble?

Parts of the market are overheated, as in every previous wave. The underlying capability is real and durable. The mistake is confusing the noise of the wave with the value of the work.

Should we wait until it settles down?

No. Your staff are already using these tools informally. The safer move is to decide deliberately what is worth doing, put practical governance around it, and start small.

We have already started. Does this still apply?

Especially then. Most organisations have plenty of AI activity and little clarity about what is working. Deciding what to scale and what to stop is what turns activity into value.