A marketing team uses a shared prompt to generate product descriptions. Team members keep saying the output is "not quite right," but they disagree with each other about what's wrong, and revisions go back and forth. What is the most appropriate first step to resolve this?

Answer: A — Agree on what a good description must contain, with one or two approved examples, and add that to the prompt.

When people can't agree on what's wrong, the root cause is vague success criteria; nobody has defined what done looks like, so Claude can't aim for it. Option (a) defines the criteria and gives examples. (b) multiplies inconsistency. (c) gambles on variations without a shared target. (d) changes tools without fixing the missing definition, which would follow them to any tool.