Five questions to answer before you start an AI pilot

Most AI pilots that disappoint were set up to fail before the first tool was chosen. Five plain questions, answered in writing, make the difference.

1. Which decision or task will change?

Start with the work, not the technology. Name the task the pilot will change, who does it today, how often, and how long it takes. If you cannot describe the task in two sentences, the pilot is not ready to start.

2. How will you know it worked?

Agree one or two measures before you begin: hours saved a week, turnaround time, error rate or customer response time. Record today's figure first. Without a baseline, every pilot looks like a success to someone and a failure to someone else.

3. What data will it touch, and who is responsible for it?

Write down which information the tool will read or create, where it is stored and who may see it. In South Africa, personal information falls under POPIA, so check that the use is lawful, that people have been told, and that a supplier outside the country handles it properly. Your Information Officer should sign this off before any customer or employee data is used.

4. Who owns the pilot?

Give the pilot one owner with the time to run it, and one sponsor who can decide what happens next. A pilot owned by a committee drifts. A pilot owned by the IT team alone often solves a problem the business does not have.

5. What happens on the last day?

Decide now what result means scale up, what means change course and what means stop. Put a date on it. A short pilot with a clear decision at the end is worth more than a long one that quietly becomes part of the furniture.

A one-page brief is enough

Answer these five questions on a single page and most of the risk disappears. If it helps, AfriTreasures Advisory reviews the brief with you, tests the numbers and gives you a written scope and fixed fee before any work begins.

← All insights