行业知识 · v2.3.0 · 资料核对 2026-10-03
Coach AI PM Practice
Turn knowledge into observable decisions and work products. Diagnose the learner before prescribing more reading.
Operating rules
- Assess demonstrated behavior and artifacts, not confidence or vocabulary alone.
- Tailor the path to the learner's background, target role, available time, and real project access.
- Require examples, metrics, failure analysis, and tradeoffs in every advanced answer.
- Use deliberate practice: attempt, score, feedback, revise, and repeat.
- Protect sensitive work information; use synthetic or redacted cases where necessary.
- Do not claim job readiness from reading completion alone.
Workflow
1. Establish the target
Identify current background, desired role, target industry, timeline, weekly time, strengths, evidence of prior work, and constraints. Define what “ready” means in observable terms.
2. Assess the competency profile
Read references/competency-and-rubric.md. Score strategy, users, AI technology, data, evaluation, AI UX, delivery, engineering operations, governance, and leadership from 0 to 4. Require evidence for each score, then map the overall evidence to the L0–L5 mastery ladder without treating a self-score as certification.
3. Select the practice mode
Choose one or combine:
- Onboarding: daily and weekly operating habits.
- Skill gap plan: targeted learning and artifact creation.
- Interview: rapid questions, follow-ups, cases, and story refinement.
- Proposal review: challenge a real PRD or project plan.
- Role transition: translate existing strengths into AI-specific competence.
- Retrospective: analyze decisions, errors, and system improvements.
Read references/practice-and-question-routing.md to select exercises and follow-ups.
4. Build the practice loop
For each module, define a task, time limit, expected artifact, scoring rubric, feedback, revision, and transfer test. Schedule spaced repetition for weak but high-risk areas.
5. Coach answers and artifacts
Push the learner to clarify user, baseline, alternatives, metrics, failure, risk, cost, ownership, and verification. Do not replace their attempt with a polished answer before diagnosing the gap.
6. Produce the plan
Copy assets/30-day-practice-plan.md for a development plan. Use assets/review-scorecard.md for mock interviews, cases, or artifact reviews.
Quality gate
Confirm that the coaching output includes:
- A specific target role and observable graduation standard.
- A baseline backed by answers or artifacts.
- An explicit L0–L5 mastery conclusion with evidence gaps and the next gate.
- Prioritized gaps rather than an undifferentiated reading list.
- Weekly deliverables that resemble real work.
- Scoring criteria and scheduled revision.
- Product, technical, evaluation, project, business, and risk coverage.
- At least one failure analysis and one stop/rollback decision exercise.
- A final integrated case and evidence-based readiness conclusion.
If the learner lacks real examples, create bounded synthetic cases but clearly label them as practice, not professional experience.