“Where Can We Use AI?” Might Be the Wrong First Question

"Where Can We Use AI?”

It is a question I hear more and more from companies. And it is easy to understand why. AI is moving incredibly fast, new capabilities appear almost every week, and leadership teams are under pressure to understand what it means for their business. There is a sense that companies need to act, experiment, and find opportunities before they fall behind.

But I think “Where can we use AI?” is often the wrong first question.

The problem is not the interest in AI. The problem is where we start. When we begin by asking where we can use a particular technology, we are already assuming that technology is the solution. We start looking across the organization for potential AI use cases before we have clearly identified which problems are worth solving.

A better starting point is much simpler: what are we trying to improve?.

Start with the problem, not the technology

When a process is slow, expensive, or heavily dependent on manual work, it is tempting to immediately look at how AI could make it more efficient. But doing so can mean skipping the most important question: why does the process require so much effort in the first place?

The underlying problem may have little to do with the task we are trying to automate. Information may be scattered across different systems. The same data may be entered several times. Responsibilities may not be clear. A process may have accumulated steps, controls and approvals over the years that made sense at one point but no longer add enough value.

And sometimes the problem is even more basic: we are doing something because we have always done it that way, without questioning whether it is still necessary.

This is why understanding how work happens today matters. Before deciding on a solution, we need to understand where time is being spent, where bottlenecks occur, which activities genuinely add value and which ones could be simplified, changed or eliminated.

Only then can we make a good decision about what comes next.

The best solution may not be AI

Once the problem is clear, the conversation about technology becomes much more useful.

In some cases, AI will absolutely be the right answer. But in others, a relatively simple automation may solve the problem. Better integration between existing systems may remove hours of manual work. A change in responsibilities may eliminate unnecessary handoffs. Simplifying a workflow may have a greater impact than introducing a new technology.

And sometimes the right answer is to top doing something altogether.

This distinction matters because there is a big difference between finding a place where AI can be used and finding a problem where AI can create meaningful value. Most organizations can probably identify dozens of potential AI use cases. That does not mean all of them deserve investment.

Starting with the business problem changes the conversation. Instead of asking, “Can AI do this?”, we can ask, “What is the best way to achieve the outcome we need?”

AI then becomes one possible answer rather than the objective itself.

Don’t automate a process that needs to be redesigned.

There is another risk in starting with technology: we may become too focused on making the current way of working faster.

Efficiency is important, but transformation should not simply be about doing the same things more quickly. Before automating a process, we should also challenge the process itself.

If we were designing it today, would we create it in the same way? Would we need all the same steps and approvals? Could some decisions happen earlier? Could information flow differently? Are there activities that could disappear completely? Where is human judgment genuinely valuable?

This is the difference between improving the current process and designing a better one.

Understanding how things work today gives us the starting point, but we also need a clear view of how they should work in the future. Once that future state is defined, we can decide what combination of process changes, automation, system integration, data, and AI will help us get there.

This is also how we approach transformation at Levare. We start by understanding the business problem, the processes behind it, and the outcome the organization is trying to achieve. From there, we can determine what needs to change and where technology can create real value.

Technology then becomes an enabler of transformation rather than a solution looking for a problem.

From a successful pilot to real impact

The same principle applies when companies move from experimenting with AI to implementing it at scale.

A technology can work perfectly well and still fail to create meaningful business value. A successful pilot proves that something can work. It does not necessarily prove it will improve how the organization operates.

For that to happen, the solution needs to fit into the way people work, connect with existing systems, have access to reliable data and solve a problem that matters enough to justify changing the way things are done.

This is why technology alone is rarely enough. Processes, people, data and systems all need to work together if an AI initiative is going to move beyond experimentation and become part of how the business actually operates.

The question, therefore, is not simply whether the technology works. It is whether the business works better because of it.

Conclusion

The real measure of success is not how many AI use cases a company identifies or how many AI tools it implements. It is the impact those changes have on the business: better decisions, less unnecessary work, a better customer or employee experience, greater speed, lower costs, or new opportunities for growth.

At Levare, we see AI as one of the tools that can help companies achieve those outcomes. It can be a powerful part of transformation, but it should not define the problem we are trying to solve.

So perhaps the next time the conversation starts with “Where can we use AI?”, it is worth asking a different question first:

“What do we need to change — and what is the best way to change it?”

Because the goal is not to use more AI. The goal is to create a better business.