People rarely bring a business analyst a problem. They bring a solution they have already picked: more scanners, a new report, an extra approval step. Sometimes the solution is right. Often it treats a symptom, and the real cause sits one step back.
Before you write a single requirement, write a problem statement. A common shape has four parts:
- The problem of what is going wrong, in plain words.
- affects who suffers from it.
- the impact of which is what it costs: time, money, customers, risk.
- a successful solution would what must be true afterwards, without naming a product.
Then test the requested solution against it. Ask "why?" until you reach something the requester cannot fix by buying the thing they asked for. If the answer is "because that is how we have always done it", you have found the need.
A problem statement names the problem and the outcome, and leaves the solution open.