The leader track 🤝 · for heads of product, founders & stakeholders · 6 x 45 min · no code to write
Thinking mode: what AI actually changes in product work and what it cannot touch, how to tell research from confabulation in your own team's documents, what a spec must contain before anyone signs it, prioritisation without theatre, metrics and the readout that admits it did not work, and how to fund an AI-assisted product org without hollowing out the craft. Each session ends with the questions to ask your PMs.
What AI changes in product
The production half of the job collapses; the commitment half does not move at all. Where that leaves your PMs, and what to stop asking them for.
🔎Leader 2 · easyThe evidence standard
Fluent research summaries are free now, so the bar moves to provenance. How to tell a real finding from a plausible one in a document you did not write.
✍️Leader 3 · mediumSpecs, decisions, accountability
What a spec must contain before it is signable, and why a drafted document with no trade-off means two people agreed to different things.
⚖️Leader 4 · mediumPrioritisation without theatre
Scoring frameworks are lenses, not answers - and AI will happily fill in inputs it does not have. How to spot a ranking built on guesses.
📉Leader 5 · mediumMetrics and the honest readout
Leading versus lagging, guardrails, and the cultural question underneath: does your org reward a readout that says the bet did not work?
👥Leader 6 · mediumThe AI-assisted product org
What to fund, what to review, what to leave alone - and how to keep the craft with the PM when the documents get cheap.
The PM track 🛠️ · for practitioners · 10 sessions x 45 min (b10: 60) · bring a real backlog item
One decision, start to finish. You are the PM on Cadence, the AI note-taker from this series - already branded, marketed and financed in the sibling courses, with nobody having decided what it builds next. Three people have asked for three different things. By b10 all three are comparable, two have not survived, and the one that did has a spec somebody can build from.
The job, and a governed setup
What you delegate, what you never delegate, what you must never paste - then meet Cadence and the three asks on your desk.
🎯PM 2 · easyFrom a vague ask to an evidenced problem
Turn "everyone is asking for it" into a problem statement someone could disagree with on the evidence - and watch two of the three asks weaken.
🧪PM 3 · mediumResearch synthesis, and the trap
Thirty snippets, clean themes in seconds - one invented, one over-generalised from a single quote. The highest-risk PM use of AI, made visible.
🌳PM 4 · mediumJobs to be done, opportunity tree
From verified themes to the job, then the tree - and the branch you chose, with the branches you cut written down beside it.
📋PM 5 · hands-onThe spec, and the scorer
Six decisions a spec must contain, five real drafts scored against them: 15, 35, 55, 70, 100. Then paste your own in.
🔀PM 6 · hands-onPrioritisation and trade-offs
Score the three asks under two lenses, watch the ranking move, flag every input the tool guessed. Then choose, and write what you gave up.
📐PM 7 · hands-onMetrics and instrumentation
The North Star, the guardrail, and the event spec written before the build - because the metric you forgot to instrument is the metric you will not have.
🧫PM 8 · hardestExperiments and honest readouts
Design the test, decide the stopping rule in advance, then write the readout that says no - the document most orgs quietly do not reward.
🗺️PM 9 · hardestRoadmap narrative and alignment
The roadmap as an argument rather than a list of dates, and the three stakeholder conversations it has to survive - including the CEO's agent.
🏁PM 10 · hardest · 60 minCapstone: one decision, end to end
Your own backlog item through all five checkpoints - evidenced problem, verified research, signable spec, chosen trade-off, instrumented metric - then the handoff.
This course owns discovery to decision and stops there on purpose. Charters, milestones, critical path, risk registers and status reports are learn-ai-project-management-with-phoebe, and b10 hands over to it explicitly. The two are a pair: product management decides what to build and why, project management gets it built. For the metrics side deeper than b7 and b8 go, see learn-metric-decomposition and learn-experimentation.
Built from PRD and discovery practice, jobs-to-be-done, opportunity-solution-tree thinking, and the research-quality literature - full source map and honest out-of-scope list in materials/official-course-map.md. The spec scorer's ladder (15 / 35 / 55 / 70 / 100) is measured rather than asserted: six real checks run over the actual text of five drafts.
Choose your path 🗺️
Two doors, one decision. Leaders never write a spec; PMs never sit through theory they have not just applied to their own backlog.
The knowledge map 🧠
Both tracks at a glance - hover a session to spotlight its concepts, click any node to jump in.