The question everyone is asking

The AI narrative is everywhere at the moment.

Magazine articles, LinkedIn posts, conferences, the national press – it’s hard to avoid, and the pace of development is clearly significant. The capabilities are real. In some areas – analysis, pattern recognition, processing large volumes of information – it’s already changing how work gets done.

Even full-length LinkedIn articles. There seems to be no end to what you can do with AI.

Against that backdrop, I’ve been asked a few times over dinner whether AI will put Mentor out of business.

After all, isn’t program execution just formulaic?

As one Chairman of a PLC once put it, “Why don’t you just give us the tick lists? Then we can do it ourselves.”

A rather quaint view – and not an uncommon one.

This from a very smart man with an impeccable track record in corporate life.

It’s a fair question – but a tad naive. And it misses where programs actually succeed or fail.

Where things actually go wrong

Because on paper, most programs look fine. The plan is clear. The numbers stack up. The work has usually been delivered to a high standard by strategy consultants- and in its own terms, it is a success. But that, on its own, is nowhere near enough.

The work has been handed over. The program still has to be delivered. The miracle has yet to occur.

And this is where things start to come unstuck.

Not dramatically at first. Just enough friction to slow things down – and then take control. A decision that doesn’t quite land. A dependency that is “under control” but not really.

Two teams who both believe they own the same issue and neither gives way. A piece of work that sits somewhere with other priorities – and always will.

None of this is hidden.

People know. They just don’t force it.

There is still a large gap between knowing what should be done – and actually doing it. That gap has been there for years. It hasn’t gone away. AI won’t close it because it sits in decisions people have to take – and the consequences that follow.

What we see when we arrive

By the time we become involved, the strategy work has already been done. The plan is in place. The thinking is sound. But the organisation is starting to struggle with execution – often having to infer quite a lot from highly compressed, high-level strategy statements.

That does mean we tend to see things when they’re not working. But after a while, the patterns become very clear – and they show up much earlier than most people expect. And they show up in remarkably similar ways, regardless of the organisation.

Not because the problem is unclear. Because a lot of detailed things have to be decided – things that can’t simply be inferred from the strategy work.

A milestone is at risk – but fixing it means admitting it won’t be met. Two parts of the organisation are working at cross purposes – but resolving it means one of them loses control.

The plan assumes capacity that simply isn’t there – but acknowledging it means resetting expectations.

These are not complicated problems. They are difficult decisions.

Because someone has to take a position, make the call, and deal with the consequences that follow.

That is where programs slow down.

Not because the problem is invisible – but because no one is prepared to force the issue.

Where AI helps – and where it doesn’t

Now introduce AI into that picture.

AI can analyse the plan. It can compare versions. It can highlight inconsistencies. It can produce better models, better reports, better decks. It can make the whole thing look sharper, more coherent, more convincing.

All of that is useful.

But it doesn’t close that gap – which can be the difference between success and ignominious failure.

Nothing is more open to misinterpretation than program requirements.

AI can also take you down the wrong path. It will infer, fill in gaps, and present something that looks coherent – but may be well wide of the mark. Unless you’re on top of it, it’s very easy to take the wrong fork in the road – and be halfway down it before you realise.

Not because the problem is unclear. Because a lot of detailed things have to be decided – things that can’t simply be inferred from the strategy work.

Because the gap isn’t analytical. It sits in decisions no one wants to take, trade-offs no one wants to own, and consequences no one wants to carry.

AI doesn’t carry consequences.

People and organisations do.

Where Mentor fits

That is why what we do is different.

We don’t arrive with large teams or people learning on the job. We work as delivery partners – and there is a difference. We’re not there to fill in forms or tick lists. That, on its own, doesn’t get the job done.

The success of what we do is dictated by the tactics we use – and they’re not captured in any framework in any meaningful detail. That’s where AI is less helpful, particularly in complex, multi-faceted situations.

Those tactics show up in how decisions are taken, how conflicts are resolved, and how the organisation is made to move.

Where this leaves us

So no – I don’t think AI will put Mentor out of business.

If anything, it will make the difference clearer. It will strip away a lot of the work that used to sit around programs – reviewing documents, building models, producing reports, and turning the same material into increasingly polished slideware.

And it is very good at those things.

But helping to shape a plan is not the same as validating it.

Plans are built on assumptions – about sequencing, dependencies, capacity, suppliers, and behaviour.

AI can help articulate those assumptions, but it can’t tell you whether they will hold under pressure. And it certainly can’t compensate for the way people actually behave when trade-offs become real and consequences start to bite.

That is where programs succeed or fail.

Used well, AI is powerful.

Used without judgement, it can be dangerous – like letting a child play with a live grenade. It looks harmless right up to the point it isn’t.

So yes – use AI. Use it hard. Use it where it is genuinely strong.

But think very carefully before trusting it to run your business-critical programs.

AI can help shape the program plan.

It still won’t make it work.

About the author

David Hilliard is founder of Mentor, specialists in strategic program execution.

You can call him on 0118 359 2444 or email david.hilliard@mentoreurope.com.