Name the problem before you name the tool

If your leadership team wants to use AI, write down the problem before you choose a tool. A named problem with a number in it gets finished. A wish to “use AI” gets a demo, then a stalled pilot.

How it usually goes wrong

A board member or a competitor mentions AI. Leadership asks for a plan. The plan is a list of products.

Six months later there are licenses, a few enthusiasts, and no answer to a simple question: what changed?

The tool was never the point. Nobody had said what it was for.

The one-sentence test

Say the problem in one sentence. It should have these parts:

• Who feels it
• What goes wrong
• How much, in hours, dollars, or days
• Who owns fixing it

Here are two examples, both illustrative.

Our two estimators spend about ten hours a week re-typing supplier quotes into the job system. The operations manager owns the fix.

Permit requests take about eleven days to answer because staff search three systems for prior cases. The community development director owns the fix.

Now compare them with “We want to use AI to be more efficient.” That sentence cannot be tested, owned, or finished.

The number test

If the sentence has no number, keep working on it. You do not need precision. You need a starting point you can measure against.

Rough is fine. “About ten hours a week” is enough to tell later whether anything improved.

Pick two or three, not twenty

Leadership teams often produce a long wish list. Choose two or three, using three questions:

1. How much time or money does it cost today?
2. Is the information involved in decent shape? If the files are a mess, fix that first.
3. Can one person own it? Problems that need six departments to agree rarely start well.

The rest can wait. A finished small project earns the right to a bigger one.

Measure before you buy

Take the baseline first. How long does the task take today? How often does it go wrong? Write it down and date it.

Without a baseline, every pilot is judged by opinion. The first disappointing week ends the effort, and a good result cannot be proven.

Questions for your leadership team

Try these at your next meeting.

• What are the two or three problems we most want to solve?
• What does each cost us today?
• Who owns each one?
• How will we know it improved?
• What would we need in place, such as clean data or a policy, before we started?

If the room cannot answer, that is a useful finding. It tells you where to spend the next few weeks.

Where this shows up in the assessment

This is sub-score 2.1, strategic intent and use cases, and 2.3, measurement and return, in the Business Direction pillar. A low grade on these two usually means the tool conversation started too early.

A short workshop can fix it. It produces a ranked list of two or three problems, each stated with a number, an owner, and a measure. If you want to see how your organization would grade, read how the assessment works.