A marketing lead notices that Claude-generated customer personas from the team's Project are almost always described as men in their thirties, even for products bought by a wide range of customers. What is the most appropriate response?

Answer: A — Review the Project's instructions and examples for skew, add guidance reflecting the real customer base, then regenerate and check.

A consistent pattern points to a cause, often in the instructions, examples or prompts, so fixing the source and rechecking stops it recurring (a). Editing each output (b) treats the symptom and the bias will keep returning. AI output is not automatically neutral (c). Abandoning personas (d) discards a useful tool instead of fixing the problem.