When AI Is the Wrong Answer to a Business Problem

A request for an AI assistant can begin with a problem that has very little to do with AI.

The pattern I encounter is straightforward: two systems define the same thing differently. Nobody has agreed what counts as a customer or a closed deal. Reports conflict, people lose trust, and an assistant is proposed to make the information easier to use.

But easier access does not settle the disagreement. A model can produce a fluent answer while the underlying definitions remain inconsistent.

In that situation, agreeing the definitions is the work that matters first.

Start with the process

I want to understand what people are trying to do before discussing the technology.

If a process follows fixed steps and the inputs are reliable, conventional automation may offer a simpler, more dependable route. If the information is incomplete or scattered, improving it may be the priority. If nobody has authority to make a decision, another tool will not resolve the ownership problem.

AI becomes worth investigating when the work involves language, variation, or judgement at a volume that makes assistance useful. Even then, someone needs to define what a better result looks like and how it will be checked.

The question is whether AI earns its place in this particular process.

Ask questions that expose the investment case

Before supporting a proposal, I ask:

  • What task or decision will change, and who handles it today?
  • What does the current process cost, including checking and rework?
  • What will we measure after the change?
  • What happens when the output is wrong, and who notices?
  • Who will own the capability a year from now?

Those questions can reveal a promising opportunity. They can also reveal that the business does not yet understand the problem well enough to select a solution.

That is a useful finding. It gives the next piece of work a purpose.

Put the alternatives on the page

When a leader wants AI and I think a simpler solution would work, I take the request seriously.

There may be a valid ambition behind it: faster service, less manual work, or a better experience for customers. My responsibility is to examine how best to achieve that ambition.

I explain what I would need to see for AI to be the right answer. Then I compare it with the simpler option, including costs, dependencies, and what each approach can deliver.

Putting the comparison in writing makes the discussion concrete. It also makes the assumptions open to challenge.

Leave room to be proved wrong

Sometimes a small test is the best way to resolve the uncertainty. Agree the measure beforehand, keep the scope contained, and decide what result would justify proceeding.

The test may support my recommendation. It may prove me wrong. Either outcome is useful if it improves the investment decision.

Our AI Opportunity and Value Assessment helps clients decide what deserves investment and why. Recommendations can include improving existing processes, investigating further, or waiting. Implementation is a separate choice.

Before committing to another AI project, talk to us about assessing the problem and comparing the available routes.

Back to Resources

Which of these sounds like your business?

Pick the closest, or send an enquiry and tell us the decision you are trying to make. We will tell you plainly whether we can help.

What happens next: we reply, then have an initial conversation. If an engagement is appropriate, a written scope sets out what you receive and what we need from you before any work begins.