A simple prompt technique that improves AI outputs by replacing guesswork with structured discovery. Learn how to use interview-style prompting with Microsoft Copilot to get better results.
A Prompt Technique That Improves Copilot Results
There’s a prompt technique I’ve migrated to Copilot Cowork now that it is available this week that gives me better results for almost any task. I previously used it on Claude Cowork and my private OpenClaw agents. It’s easy, but it’s worth sharing.
“Interview me about what I want to accomplish. Don’t stop until you really understand my goals.”
Instead of me writing a massive paragraph of content upfront while simultaneously wondering if I am steering into my biases and creating an overly agreeable assistant. Copilot asks me questions. Some are typical, some are not ones I would have expected. I try to thoughtfully answer. By the end of the conversation, it has everything it needs to do the work and I haven’t had to figure out how to structure a prompt that covers all the angles I’d otherwise miss. This results in much better responses, actions, content, etc.
Why Front-Loading Context Fails
The standard advice for AI prompting is: Give it as much context as possible. Be specific. Include the goal, the constraints, the audience, and the format.
In theory, in practice, most people – including me – get a little short with what they send in that initial content package. You write a detailed prompt, the output misses something obvious, and you realize you forgot to mention a key detail that only became obvious after you read the draft, because the draft is not the direction you intended.
The interview technique really helps with this. Each question forces you to articulate something you might not have included in a raw prompt. By question four or five, you’ve surfaced constraints and requirements you didn’t know you had. One way it helps me is when I know the outcome I want, but I haven’t really sat and thought it through enough to know all the steps. This kind of AI strategy therapy through interrogation helps me clarify the request, steps, and goal in my own mind. Ultimately, that helps support the follow-through.
When To Use It
This works best when the task is complex enough that you’re not sure what the right framing is. Primarily in content/communication creation, like strategy documents, campaign briefs, planning outlines, anything where the structure matters as much as the content. However, it’s also been effective in helping me program new skills or think through complex workflow logic.
For simple tasks, just prompt normally, but when you’re about to write three paragraphs of context, and you’re not sure you’re covering everything, try the interview instead. You’ll get better output, and you’ll learn something about the problem in the process.
Better yet, create yourself an interview agent. Train it so it improves and bends towards your way of work. Telling Copilot to extract context rather than assume it, and forcing you to think through the problem by answering questions rather than by trying to write a perfect brief.
Most of the time, the questions are better than the prompt you would have written.
Key Takeaway
Instead of trying to write the perfect prompt upfront, let Microsoft Copilot guide you there. Using an interview-style prompt helps surface hidden goals, constraints, and assumptions, leading to clearer thinking and significantly better AI-generated outcomes.
Turn Better Prompts Into Real AI Outcomes
Move beyond experimentation. Identify high-impact AI use cases and build a clear roadmap to scale what works across your organization.
