AI hype doesn't pay the bills

CIOs are pouring more of their budgets into AI. Yet many still cannot clearly prove the return. That was the thrust of a recent CIO Dive article, and it gets to the heart of a problem I personally see all the time.

Many companies are buying AI because it feels like something they need to have – rather than because they have a clear understanding of what results it will deliver.

Start with the commercial outcome

AI gets discussed in grand terms, with talk of transformation and reinventing the future of work. Some of that may be true – in time. But for leaders who own budgets and delivery, the first question is simpler: what, exactly, is this going to do for the business in a tangible way?

If the answer is vague, the project is in trouble. A useful AI project should be tied to hard commercial outcomes from the start: higher revenue or lower costs.

How consultancy drift loses the value

But this is, sadly, where many consultancies don't hit the mark. They sell AI as the next hot thing, then let the work drift into something much more amorphous. The goal posts move and the value gets lost.

I have a term for this: consultancy drift.

So, the difference between a useful AI programme and an expensive science experiment is not technical brilliance. It is whether it delivers measurable commercial results in a clearly defined period. I am a big believer in tech, and in AI specifically, when it is applied properly.

But in practice, some of the best results come from the more boring parts of the workflow, whether that's process automation, better routing, less manual handling, or faster reporting.

This reality is coming to bite for many companies, and I cannot shake the feeling that it's harming the standing of the consultancy industry.

For example, McKinsey found that 78% of organisations now use AI in at least one business function, up sharply from 55% the year before. Yet only a much smaller group report a material impact on the bottom line. That gap matters. It shows that adoption alone is not the win. Delivery is the win.

Results only count if they last

The other problem is sustainability. A project can look great in week one after completion and, yet, still fail six months later. What happens? Well, the teams revert to old habits and, as a result, the promised savings start to disappear.

That is why onboarding, training, and centralised follow-through matter just as much as the original build. If behavioural change does not stick, neither do the numbers.

At Trend-Setters Consulting, we've seen this play out time and time again in enterprise environments. We have a strong track record of helping clients reduce their dependence on print and paper processes.

But we don't stop there when the change is in place. We stick around monitoring KPIs for months to ensure the change sticks. And the same needs to be done for AI projects.

How to test a consultancy's AI claims

So, as a client, how do you know if you're going to beat the odds and actually integrate AI in a way that delivers good return?

It's difficult, no doubt, but when a consultancy says it has AI expertise, the right response is simple: show me the case studies, the cost, and most importantly, what was actually delivered.

That is the real test. Not whether a firm has a massive brand, large team, or can talk well about AI. It needs to all come back to whether they can turn AI into revenue, savings, and results that hold.

That's what separates pretty decks from genuine AI strategies; hype from results.