The 10 Minutes That Change Your Week - PM AI Habit Stack
ProjectorPM Signal · May 2026
Most PMs who “tried AI” spent an afternoon asking it to write emails and went back to their inbox when the outputs weren’t quite right. That experience is so common it’s almost universal. And it’s almost entirely a problem with the entry point, not the tool.
The Signal
The PMs who get the most from AI aren’t the most technically capable. They’re the ones who started using it on real work, consistently, until they knew it well enough to trust it. That’s a habit, not a skill.
Habits are built with small, concrete starting points. Not courses. Not experimentation sprints. Not afternoon workshops. A recurring, bounded task that produces something you’ll actually use - done again and again until it’s just how you work.
The most effective entry point I’ve found: turn a vague requirement into acceptance criteria. It’s the right size (10 minutes), the output is immediately useful (you’d produce it anyway), and it’s immediately evaluable (you can tell whether it’s good enough to use without needing a specialist’s eye).
The Noise
“I need to do a proper AI training course before I can use this effectively.”
The course doesn’t give you the habit. Only doing the work gives you the habit. And the most valuable thing a course teaches you - what AI needs from you to produce good output - you can only really learn by running prompts on real data and seeing what comes back.
Skip the course. Start with one real task.
The Tool
Here’s the Day 1 prompt from the PM AI Habit Stack. Take the vaguest requirement currently in your backlog - the one you’ve been avoiding because it’s written too broadly to act on. Paste it in. Run this prompt.
You are a senior project manager helping to define clear, testable acceptance criteria.
I have a requirement that's written too broadly to work from. Using only the information I provide, convert it into a set of acceptance criteria that:
- Are specific and measurable (no ambiguous language like "user-friendly" or "fast")
- Are written in plain language - no jargon unless it's already in the requirement
- Follow the format: "Given [context] / When [action] / Then [outcome]" where useful, or a numbered checklist where that's cleaner
- Flag any assumptions you've had to make - mark them clearly as [ASSUMPTION: ...]
- Flag any gaps where I need to provide more information - mark them as [CLARIFICATION NEEDED: ...]
Do NOT invent requirements that aren't implied by what I've given you.
VAGUE REQUIREMENT:
[Paste the requirement here]
ADDITIONAL CONTEXT (stakeholder expectations, constraints, what's been discussed verbally):
[Paste any relevant context here - or write "None provided"]If the output is weak, look at what you pasted in. AI calibrates to the specificity of what you give it. A three-line description produces a three-line-depth output. Add more context and run it again. That adjustment - not the first output, but the act of figuring out what context was missing - is where the learning happens.
The Full Stack
7 days of foundation prompts - each one producing a real artefact from your real project data. Free, email only.
[Get Week 1 free → projectorpm.xyz/pm-ai-habit-stack]
P.S. The full 30-day stack covers requirements, analysis, stakeholder communications, and building your own prompt library. Each day: one task, one artefact, ten minutes. By Day 30 you’ll have a documented working practice - not just a habit. $29 for the lot.
© 2026 ProjectorPM · projectorpm.xyz
Yes, AI helped me to write this :)