Where to Start with AI: Use Cases Before Tools

Imagine a management meeting where someone says, "We need to do something with AI." Everyone agrees. Then the room goes quiet, because nobody can say what the something should be.

The problem: "Everyone is talking about AI and we do not know where to begin"

The usual first move is to pick a tool. A demo looks impressive, a colleague recommends a product, and a subscription is bought. Then the team asks what to use it for, and the answer is vague.

Meanwhile the real questions go unasked. Which work in our company is slow, repetitive or error-prone? And is the information that work depends on even in a usable state?

What it costs

• Tools without a purpose: licences and pilots that nobody uses once the novelty fades.
• The wrong first project: effort goes to what is exciting rather than what matters, and an early disappointment makes the next proposal harder to approve.
• Hidden blockers: data scattered across files, inboxes and old systems only comes to light halfway through, when changing course is expensive.
• Scattered effort: each department experiments on its own, and nothing adds up to a direction.

The turning point: start with the business, not the tool

The better order is the reverse. Decide what the business needs, check what it already has, and choose the tools last.

1. Name the use cases. Sit with the people who do the work and list where time is lost: repeated questions, manual copying between systems, searching for documents, routine approvals.
2. Check readiness. For each use case, ask whether the data exists, whether it is accurate, and whether the systems involved can be connected.
3. Rank them. Weigh each one by its value to the business and by how ready it is. Valuable and ready goes first. Valuable but not ready needs preparation first.
4. Write the plan. A short sequence that says what to do first, what to prepare, what to leave for later, and who owns each step.
5. Decide build, buy or partner. Only now, when you know what you need, compare the ways of getting it.

What changes

This approach produces a prioritised plan the management team can act on. It is not a list of ideas. It is an order of work, with the reasons behind it and the blockers named.

The first project is chosen on evidence. Budget decisions become easier because every step has a purpose. And the tool question, which felt urgent at the start, turns into a much smaller decision.

A tool answers a question. First make sure you are asking the right one.

What this means for your company

Before your next AI purchase, ask two things. Which business problem does this solve? And have we checked that our data and systems can support it? If the answers are unclear, a short readiness exercise is a sensible step before committing budget.

This is the work covered by my AI readiness and strategy service: use-case workshops, a data and systems readiness review, a prioritised plan, and a build, buy or partner recommendation. It is the same approach I apply in the AI solutions I have delivered for large enterprises and public-sector organisations over the past year.

If this sounds familiar and you would like to talk it through, the WhatsApp button on this page is the quickest way to reach me.