AI tools and prompting interview questions

I give the assistant a role, the relevant facts, clear limits such as length and tone, and the exact format I want back. For example, for a support reply I include the customer's issue, our policy and a 120-word limit. If the first answer misses, I give specific feedback rather than starting over.

I treat it as a draft and verify every figure, name and claim against a source I can read myself. For formulas or code, I test them on cases where I already know the answer. Anything I cannot confirm, I remove.

Customer personal data, passwords or keys, confidential pay or contract details, and internal plans, unless my company has approved a specific tool for that data. When I need help with such material, I replace real details with placeholders like [Customer A]. I also follow the employer's written AI policy.

A good answer names a real case: for example, an assistant summarised a policy and stated a 30-day refund window when ours was 14 days. Explain how you caught it, such as comparing with the source document, and what you changed afterwards, like always pasting the policy text into the prompt.

When the decision carries serious risk for a person, such as medical, legal, financial or hiring decisions, and when there is no way to verify the output. AI can help prepare information, but a qualified person must review and decide. I also avoid it where the company policy does not allow the data involved.

I would create shared prompt templates for recurring tasks, each with a role, examples and a fixed output format. I would also agree on a review checklist, such as checking facts, tone and privacy before anything goes out. Templates get updated when someone finds a better version.