What belongs in an AI PM portfolio?
Show the user task, evidence, non-AI baseline, AI rationale, evaluation set, rubric, failure handling, metrics, personal contribution, and project limitations.
Loading...
Loading...
Check problem evidence, AI choices, evals, integrity, metrics, and presentation. Receive up to three priority revisions—not a fabricated hiring verdict.
Start the 12-point checkFully public criteria
Individual answers are not sent to analytics
Prioritizes edits; never predicts offers
Do you support the user problem with interviews, current workflows, public data, or reviewable observations?
Problem evidence
Do you define the target user, core task, usage context, and explicit non-goals?
Problem evidence
Do you compare AI with human, rules, search, or template-based baselines?
AI decision
Do you explain model, retrieval, or tool choices and the quality, latency, and cost tradeoffs?
AI decision
Do you have an evaluation set covering typical, edge, and high-risk inputs?
Evaluation
Do you include a reusable rubric, failure taxonomy, baseline, and regression result?
Evaluation
Do you distinguish your contribution, team work, borrowed assets, and project limitations?
Integrity
Is simulated work labeled clearly without invented users, revenue, launches, or employer experience?
Integrity
Do you define task quality, user behavior, business value, cost, and safety metrics?
Metrics and launch
Do you specify rollout scope, launch gates, stop conditions, human takeover, and rollback?
Metrics and launch
Can you explain context, judgment, solution, result, limitations, and next step in 5–10 minutes?
Communication
Are links, prototypes, tables, and attachments readable, accessible, and free of sensitive information?
Communication
After completion, the browser stores only total score, result band, weak dimensions, and completion time. Analytics events exclude individual answers; providers may still process IP, browser, device, or session data. Read the Privacy Policy.
Show the user task, evidence, non-AI baseline, AI rationale, evaluation set, rubric, failure handling, metrics, personal contribution, and project limitations.
Yes, if simulated work is labeled clearly. Show research notes, public data, prototype tests, eval sets, failure samples, and decisions without inventing users, revenue, or launch results.
No. This is a transparent, self-reported practice tool for prioritizing revisions, not a standardized assessment or hiring prediction.