Your AI Prompts Are Too Vague: A PM's Guide to Getting Better Results

You stare at the screen. The AI-generated user story is a mess of generic business jargon. It’s technically correct, but completely useless. It has no soul, no specific user pain, and you’ll spend more time rewriting it than you would have starting from scratch.

This isn't the AI's fault. It's yours.

We treat generative AI like a lazy intern. We give it vague, one-sentence commands and then get frustrated by the mediocre output. To get great results, you need to write a prompt with the same clarity and context you’d put into a creative brief or a project spec. A better prompt is a better product.

The Anatomy of a Powerful PM Prompt

Stop using AI like a search engine. Start using it like a junior team member who is brilliant but has zero context about your job. A powerful prompt gives the AI a role, a goal, and guardrails.

I use the CRAFT framework to structure my requests.

• Context: Provide the background. Who is the user? What is the product? What business goal are we trying to achieve? Paste in the user persona or relevant data if you can.
• Role: Tell the AI who to be. "Act as a senior product manager," "Act as a skeptical CFO," or "Act as a conversion rate optimization specialist." This primes the model to use the right lens and vocabulary.
• Action: What, specifically, do you want it to do? Use strong verbs. "Write," "Critique," "Brainstorm," "Summarize," "Rephrase," "Generate a list of."
• Format: How do you want the output delivered? "In a markdown table," "As a user story with the format As a [persona], I want [action], so that [outcome]," "In bullet points," "As a 300-word email."
• Tone: Define the voice. "Professional