菜单
Turn scattered AI ambition into an evidence-aware product strategy thesis with choices, boundaries, outcomes, assumptions, and next bets. Use when direction or non-goals are unclear.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/deanpeters/ai-product-operating-model-skills --skill aipom-strategy-thesis-advisor命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
基于 SOC 职业分类
正在显示 SKILL.md
| name | aipom-strategy-thesis-advisor |
| description | Turn scattered AI ambition into an evidence-aware product strategy thesis with choices, boundaries, outcomes, assumptions, and next bets. Use when direction or non-goals are unclear. |
| type | interactive |
| category | strategy-and-economic-outcomes |
| phase | 1 |
| status | active |
| operating_level | ["organization","portfolio","product-team"] |
| audience | ["Head of Product","CPO","CTO","VP Product","Product Operations","Product Manager","Design","Engineering","Data","Finance","AI Governance"] |
| best_for | ["Converting broad AI ambition into strategic choices","Aligning leaders on where AI should and should not be used","Connecting proposed AI bets to customer and economic outcomes","Revising an AI strategy after evidence or market conditions change"] |
| evidence_required | ["Current product and business strategy","Customer or end-user evidence","Portfolio and investment evidence","Relevant product, workflow, data, evaluation, and governance constraints","Results from existing AI experiments where available"] |
| produces | ["AI product strategy thesis","Strategic choices and non-goals","Evidence and assumption ledger","Portfolio implications and next learning bets"] |
| assessment_questions | ["STR-01","STR-02","STR-03","STR-04","STR-05","POR-05"] |
| maturity_move | {"from":"emerging","to":"repeatable"} |
| estimated_time | 45-75 min |
| group_size | 2-8 |
| depends_on | [] |
| combine_with | ["aipom-bet-charter","aipom-operating-model-roadmap","aipom-outcome-value-map","aipom-economic-case-builder"] |
| sources | [] |
Use this advisor to turn broad AI ambition, disconnected initiatives, or executive pressure into a testable AI product strategy thesis. The thesis explains where the organization believes AI can create distinctive value, which customer and economic outcomes should change, what it will not pursue, which conditions must be true, and what it will learn next.
The advisor changes strategy from a collection of technology intentions into an evidence-aware set of choices that can guide portfolio decisions.
“Use AI everywhere,” “be AI-first,” and “add copilots to the roadmap” are directions of travel, not strategies. They do not tell teams which problems deserve attention, how value will be created, what tradeoffs are acceptable, or when to stop.
A useful thesis helps leaders:
A documented thesis is not proof that leaders use it. Evidence of maturity includes investment, sequencing, and stop decisions changed by the thesis.
Use this advisor when the organization has many AI ideas but weak coherence, when leaders disagree about the role of AI, when a portfolio is dominated by vendor-led experiments, or when an existing thesis no longer fits current evidence.
Do not use it to select a model, write a vendor comparison, or generate an executive slogan. If the immediate question concerns one initiative, use a bet charter or initiative-readiness review after establishing the broader strategic context.
The final strategy packet contains:
Include the leader accountable for product strategy, relevant product and technology leaders, and people who understand customers, economics, delivery constraints, data, risk, and governance.
Do not let the session represent only executive aspiration. Bring practitioner, customer, operational, or governance evidence proportionate to the decisions being made.
Use whatever reliable material is available:
Separate evidence from leadership belief. Missing evidence limits confidence but should not prevent a best-guess thesis with labeled assumptions.
Extract the organization, strategic scope, planning horizon, existing strategy, customer and economic evidence, current bets, constraints, decisions already made, and unresolved disagreements. Do not ask for information already present.
Clarify which decision the thesis must guide. Examples include narrowing an AI portfolio, choosing a customer-value arena, balancing product and internal workflow bets, or identifying what not to fund.
Determine whether the main problem is:
Focus the conversation on the most consequential gap rather than filling every template field.
Work through the choices that matter for this context:
Write the important belief chain:
If we apply AI to this arena, then this system or workflow behavior will change, enabling this customer or operator behavior, producing this product outcome, which should influence this economic or organizational outcome.
Mark each link as supported, partially supported, or assumed.
Compare the proposed direction with plausible alternatives, including non-AI options. Ask what the organization will stop, defer, or decline if it chooses this thesis.
Use this structure:
We will focus AI investment on [chosen arena] for [customers or users] because [evidence-backed insight]. We believe [AI-enabled change] can produce [customer/product outcome] and contribute to [economic or organizational outcome] when [required conditions] are true. We will differentiate through [advantage]. We will not [non-goals or boundaries]. Over [horizon], we will test [next bets], and we will revise, pause, or stop when [decision triggers] occur.
Check whether the thesis could actually reject an attractive idea. If it cannot, it is too broad to guide a portfolio.
Name the bets to explore, fund, pause, or stop; the owner of each decision; the evidence expected; and the thesis review trigger. The AI facilitator may recommend options but cannot make the strategy decision.
Ask one consequential question at a time. A useful sequence is:
Skip or adapt questions already answered.
Build an initial evidence and assumption ledger from supplied material. Draft a provisional thesis, identify only the gaps that could change it, and ask for those gaps in priority order.
Draft a thesis using labeled assumptions, state confidence, and identify the smallest evidence-producing actions needed before major investment. Do not convert an unsupported aspiration into a maturity score above 2 — Emerging.
At meaningful choices, present numbered options with fit and tradeoffs. Accept combined or custom selections and synthesize them.
Use the following routing logic:
When presenting strategic posture options, use context-specific versions of:
Recommend a posture only when the evidence supports it.
Complete the advisor when the output includes:
Do not finish with a slogan or a generic invitation to brainstorm more ideas.
A strategy defines where to focus, how value may be created, and what not to do. A list of AI use cases is not a strategy.
Begin with customers, decisions, workflows, outcomes, and constraints. Choose AI only when it improves the value logic relative to alternatives.
AI capability matters only if it changes system or workflow behavior in a way that changes customer or operator behavior and produces meaningful outcomes.
Non-goals protect attention, capital, trust, and learning capacity. Boundaries should be concrete enough to reject work.
Treat strategy as a coherent belief supported and revised by evidence, not as an eternal statement defended from inconvenient learning.
aipom-bet-charter to convert a chosen strategic bet into an owned investment hypothesisaipom-operating-model-roadmap to sequence organizational interventionsaipom-outcome-value-map to deepen the causal connection between behavior and economicsaipom-economic-case-builder to test financial and operational assumptionsThis skill is an original AIPOM synthesis of product-strategy, evidence, and portfolio-choice practices. It does not claim a named external methodology.