Think first. Then digitalise.

Artificial intelligence is supposed to save us time and money. Right now, for many businesses, it is doing one thing above all. Costing them even more money.

Licences, applications, infrastructure and a constant stream of new possibilities. The final bill could be considerably higher than expected.

That does not surprise me much.

We have seen this before. A new technology promises to make everything better at first. And in the end we have the same problems as before. Just with another piece of software.

Faster is not automatically better.

Imagine a business where the same information is entered three times, decisions pass through five departments and employees spend a sizeable part of their time looking for things or asking questions.

Then along comes AI.

Suddenly, emails are written faster, meeting minutes are produced automatically and presentations are put together in seconds.

Brilliant.

But why is the information still being entered three times? Why do we need five approvals? And who actually reads all these new presentations?

Perhaps we should get rid of the unnecessary work before we automate it.

That would at least be worth a try.

Two hours saved. Now what?

If a task used to take an employee three hours and now takes just one with AI, two hours have been saved.

But that does not automatically mean paying for two fewer hours.

Perhaps they can now look after customers better. Perhaps projects get finished sooner. Perhaps the business will not need to hire another person in future.

All very good.

But you should know beforehand what you are actually trying to achieve.

Otherwise, we end up with impressive productivity figures, extra software costs and the same organisation as before.

Saving time is not the same as saving money.

You can roll out technology. Not people.

For me, that sentence gets to the heart of a major problem.

Businesses buy software, arrange training and then wonder why their employees are still working exactly as before.

It is not really that surprising.

Someone who has worked a certain way for years will not change their habits simply because management has bought a new application.

People need to understand why things are changing. What will be different. What will disappear. And, above all, what will be better for them.

That also means listening.

The employees who deal with cumbersome processes every day often know exactly where things get stuck.

We just need to ask them.

Preferably before ordering the new software.

Perhaps we do not need AI at all.

A thought some technology providers would probably rather not entertain.

Not every problem calls for artificial intelligence.

Sometimes a sensible process is enough. Clear responsibility. Software that is already there but nobody uses properly.

And occasionally it would be progress if two departments simply talked to each other.

Of course, that is less exciting than the latest AI application.

But it might be considerably cheaper.

I am certainly not against AI. Quite the opposite. The possibilities are enormous, and many businesses would be unwise to ignore them.

But we should start with the task, not the technology.

What should get better? What is wasting our time today? What do we want to keep? And what can we finally stop doing?

Only then should we ask which tools will actually help.

And what do we do with the time we save?

If AI takes on work but businesses leave their existing processes and staffing structures unchanged, the cost savings they hoped for may never materialise.

That makes sense from a business perspective.

Even so, my first question would not be how many employees we can cut.

I am more interested in what we could do better with those people.

Look after customers? Preserve knowledge? Develop new areas of business? Tackle tasks nobody has had time for?

Of course, it may also become necessary to cut jobs or redistribute work. There is no point pretending otherwise.

But anyone who sees AI purely as a way to reduce headcount may be missing the more interesting opportunities.

AI does not make a bad process better.

In the end, it has to add up.

I believe AI will fundamentally change our businesses.

But the most successful will probably not be those that buy as many applications as possible.

They will be the ones that question their processes, bring their people with them and know exactly where their money is going.

Because what matters is not everything AI can do. It is what your business can do better as a result.