Are We About to Repeat the Same Mistakes With AI?

Are we about to make exactly the same mistakes with AI that we made with big technology?

For decades, organisations have invested millions in ERP, HRIS, shared services, self-service and automation. Yet how many genuinely delivered the ROI promised in the original business case?

There’s a well-worn saying, often attributed to Einstein: “Insanity is doing the same thing over and over again and expecting different results.” Whether Einstein actually said it or not, it feels particularly relevant right now.

Because too often, we implemented the technology but didn't fundamentally redesign the work.

We automated individual tasks instead of redesigning the end-to-end process. We introduced self-service but didn't remove the work it was supposed to replace. We digitised processes but retained the same hand-offs, checking, controls and organisational structures around them.

And when technology released capacity, we weren't always clear about where that capacity actually went.

Now we have AI

The technology may be revolutionary. But are we?

Putting AI into a badly designed process doesn't transform it. It may simply make a bad process happen faster.

An HR transaction rarely starts and finishes in HR. It can touch Finance, Payroll, Pensions, Benefits, external providers, managers and employees. Saving ten minutes in one part of that chain isn't ROI if we create another ten minutes of checking, correction or administration somewhere else.

And if AI releases 20% of someone's capacity, what happens to it? Is it removed? Reinvested? Used to improve service? Or does it simply disappear into an already busy working day?

Not prompts. Outcomes.

AI gives us an extraordinary opportunity to rethink how work gets done. But that means being prepared to redesign processes end-to-end, build governance into them, challenge organisational structures and roles, remove redundant activity and measure what actually changes in production.

Not prompts. Not licences. Not numbers of agents. OUTCOMES.

We've had decades to learn these lessons from major technology transformation. So here's the question I keep coming back to: How, as intelligent humans, do we make sure we don't use the most intelligent technology we've ever had to repeat exactly the same transformation mistakes we've been making for years?

AI adoption isn't AI transformation. And AI transformation isn't AI ROI.

What do you think – have we actually learned the lessons from the last generation of technology transformation?

Related insight

The 30% Rehire Problem