Doing more isn’t the same as creating more value

We’ve never had this much capacity to get things done. Tasks that used to take hours can be automated, we can analyse in minutes information that would have taken days of work, and the distance between imagining something and actually being able to do it has shrunk in an extraordinary way.

With AI, that feeling has accelerated even further. It’s wonderful news, but it has a less obvious consequence: something being possible, fast or cheap doesn’t mean it’s worth doing.

For a long time, a large part of technology decisions came down to whether a company could afford a given capability. Today, increasingly, the interesting question is a different one: now that we can do so much more, what actually deserves our attention?

Because technology can increase a company’s capacity without necessarily increasing its value. And confusing the two can turn out to be fairly expensive.

The hour we just saved

Picture a process that takes ten hours of work a week, and a tool that lets you do it in two. The case looks simple: we’ve saved eight hours a week, and we can even work out what that’s worth.

Except the money hasn’t shown up yet.

The person is still working the same hours and earning the same salary. What we have now is eight available hours to do something else with. If that capacity was limiting growth, the impact can be enormous: maybe we can now take on more business without adding structure. But if those eight hours simply dissolve into the working day, we’ll have made a task much better and the P&L barely at all.

This isn’t an argument against automation. Quite the opposite. It’s a reason to be more demanding about what we expect from it, because freeing up capacity and creating value aren’t exactly the same thing.

Between the two sits a business decision that technology can’t make for us: what we’re going to do with the capacity we’ve just freed up.

Starting with the tool changes the question

When a new technology appears, it’s very easy to start by thinking about what it lets you do. With AI, this is happening constantly: we see a surprising capability and immediately try to find somewhere to apply it.

It’s understandable. It’s also a fairly effective way to end up with lots of interesting projects and few results that actually matter.

If I start by asking where I can use AI, I’ll find opportunities to use AI. If I start by asking what’s stopping my business from achieving a particular result, I might find a problem AI is an extraordinary tool for. Or I might not.

And that second possibility matters.

We don’t start with the data. We start with the decision we want to improve.

It’s not just a semantic difference. It changes the order of reasoning: technology stops being the project and becomes a possible answer to a problem we’ve already understood.

Something similar happens with processes. There are slow, manual jobs that look like obvious candidates for automation, but before doing it, it’s worth asking a slightly unwelcome question: why do we do this in the first place?

Sometimes there’s a good reason, and technology will let us do it much better. Other times we discover we’re spending time on something that made sense at some point and adds fairly little value today. Automating it would deliver a flawless improvement — except we’d have invested in doing much better something we probably didn’t need to be doing at all.

This idea isn’t new. What’s changing is the scale. When automating was expensive, the cost itself forced selectivity. You had to justify fairly well why it was worth intervening. AI is lowering that barrier, and the cheaper doing becomes, the more important it is to choose well what we do.

Working doesn’t yet mean it’s creating value

That’s also why it’s worth being careful with pilots. A trial can perfectly well demonstrate that a technology works, and still not tell us very much yet about the impact it will have on the business.

Back to our eight hours. If the pilot shows we can free them up, we already know something important: the technology does what we expected. Now we need to know what will happen to that capacity once we roll the solution out into how the business actually runs.

If those hours let us absorb more business without hiring, the link to results is fairly clear. If nothing else changes, the benefit will be far harder to find.

Until we resolve that second part, what we have is a possibility, not yet a return.

This distinction helps explain why a company can rack up successful pilots and still struggle to explain what’s actually changed in the business. It may not have chosen the wrong technology. It may simply have assumed that demonstrating capability was the same as demonstrating value.

Technology also amplifies what isn’t working

There’s another reason I prefer to get to the technology after understanding the business. A tool can massively speed up a way of working without questioning it, and depending on what it finds underneath, that can be an extraordinary advantage or exactly the opposite.

Back to our eight hours for a moment. If the company knows what capacity it needs to grow, has a way of working that functions, and people can make the decisions that are theirs to make, freeing those hours up can have a multiplying effect. It might let the business absorb more work without adding structure, and, this time, actually turn technology into productivity and profitability.

But the extra capacity doesn’t always land in an organization prepared to use it.

If a process generates errors, doing it faster doesn’t fix the problem. And if a company already struggles to absorb what’s happening today, dramatically increasing its capacity to act can end up paradoxically reducing its capacity to react. Suddenly more things are happening, and happening sooner, but the organization that has to interpret them, decide, and respond is still the same one.

That’s where the conversation about AI starts to look a lot like a conversation about the company itself.

Because before asking how much we can speed up, it’s worth understanding what we’re speeding up, and whether the organization is ready to absorb it. Technology can extend our capacity to grow, but it doesn’t build the foundations that make that possible on its own.

That’s why a technology transformation often begins well before choosing any technology. It begins by understanding what company we want to build, how we intend to make money, and what has to work for that to happen. Only then can we ask where technology can multiply our capacity to do it.

Order matters.

A prepared company can use AI to do more without needing its structure to grow at the same pace. One that isn’t prepared yet can use the exact same technology to generate more activity, at greater speed, on problems that remain unsolved.

When doing stops being the hard part

For years, a large part of competitive advantage lay in being able to do things others couldn’t. Technology is narrowing that gap in many areas, and capabilities that recently required significant investment are becoming available to many more companies.

That doesn’t make technology less important. It shifts where part of the value sits.

When doing becomes easier, choosing well what we do, and building an organization able to absorb it, becomes more important. And that’s where the conversation stops being purely technological.

AI can multiply capacity, but someone still has to decide where to point it and what outcome we expect from it. The company still needs to turn that new capacity into business, without growth in activity also becoming growth in disorder.

That’s why I’m considerably less interested in how many AI use cases a company can find than in what it expects to get from them. The list of possibilities can grow very fast. The P&L, not necessarily.

And maybe that’s the simplest test for any technology transformation: not how much more we’re able to do, but what business outcome we’re able to change because of it.