Stage-Gate Review Deck Builder

How to use it

1. 1Copy the prompt into ChatGPT, Claude or Gemini.
2. 2Fill the bracketed inputs (project, gate, criteria, evidence, ask).
3. 3Answer its clarifying questions about criteria and evidence.
4. 4Iterate: ask it to pressure-test the recommendation like a hostile gatekeeper.

See it in action

Slide 1 — Gate question: Should Nordvik commit pilot budget to the drinkable skyr line, hold pending cost resolution, or stop?

Slide 2 — Evidence vs criteria:
- Consumer demand: MET — concept screen cleared the internal action standard, strongest with younger protein-seeking shoppers.
- Cost structure: NOT MET — current co-packer quote leaves margin below threshold; two levers unexplored (bottle spec, batch size).
- Chilled logistics fit: PARTIAL — fits existing lanes; shelf life at the distribution edge untested.

Slide 3 — Open risks: co-packer cost (owner: ops lead; mitigation: renegotiate bottle spec, deadline in six weeks); edge-of-network shelf life (owner: R&D; mitigation: transit trial).

Slide 4 — Resource ask: pilot production budget, insights support for a two-market test, and a named decision date.

Slide 5 — Recommendation: HOLD for six weeks. Advance only if the cost gap closes through the two named levers; kill if margin stays below threshold. Rationale: demand evidence is strong enough to protect, cost evidence is not strong enough to spend against.

Want me to draft the talk track for the recommendation slide?

The prompt

Here's the start of the prompt. Download the free bundle for the full, ready-to-paste version — plus the installable Claude Skill and Custom-GPT instructions.

# Role
You are a senior innovation portfolio director who has chaired hundreds of stage-gate meetings for food and beverage companies. You have no patience for progress theater: a gate deck exists to answer one question — advance, hold, or kill — and every slide must serve that decision.

# Context I'll provide
- Project and category: [PROJECT]
- The gate being faced: [GATE e.g. gate 2 business case, gate 3 development complete]
- Gate criteria: [CRITERIA — what this gate is judged on]
- Evidence gathered so far: [EVIDENCE — findings, results, quotes, numbers you have]
- Resource ask for the next stage: [ASK]
- Known risks or open issues (optional): [RISKS]

# Your task

Frequently asked questions

What is a stage-gate review in product development?

A stage-gate review is a formal checkpoint where a committee decides whether a project advances to the next development stage, holds for more evidence, or stops. Each gate has criteria the project must meet. The deck's job is to map evidence against those criteria honestly and land on a recommendation — advance, hold with conditions, or kill — that the committee can act on in one meeting.

How is this different from the Prototype Test Plan Builder?

The Prototype Test Plan Builder designs the experiments that create evidence between gates. This skill takes whatever evidence exists on gate day and structures it into a decision-ready deck — criteria met or not, risks, ask, recommendation. One plans the work; the other presents the verdict. Most projects use both, several weeks apart.

What if my evidence is incomplete at the gate?

That is the most common situation, and the deck is built for it. Gaps are graded 'no evidence yet' rather than papered over, and the recommendation adjusts — usually a hold with named conditions and a resolution date. Gatekeepers respect an honest gap far more than activity dressed up as proof, and a conditional hold protects the project's credibility.

Which AI models does this skill work with?

Any capable chat model — ChatGPT, Claude, or Google Gemini — handles it as written; nothing in the prompt is model-specific. If your innovation team runs gates on a fixed calendar, save it as a Custom GPT or a Claude Skill so every project lead builds gate decks with the same structure and the same honesty standards.