Product Strategy / Vision Beginner

35 ChatGPT prompts for startup founders to write a clear product vision document

A great product vision document does one job: it makes a 50-person team wake up tomorrow and pick the same fight. Most startups skip it, draft it badly, or write something that sounds like a press release and dies in a Google Doc nobody reopens. I wrote this guide because I got tired of watching early-stage founders try to “figure out the vision later” while their team quietly built four different products.

If you’re searching for ChatGPT prompts for product vision document work, you already know the bottleneck. The bottleneck isn’t typing. It’s clarity. So below I’m handing you 35 prompts I actually use with founders, organized the way I’d coach a first-time CEO through the document in 14 days. Each prompt is multi-line, has an example output, and has pro tips so you don’t ship generic mush. Pull them into ChatGPT, Claude, Gemini, or your favorite model, and edit in your own scars.

TL;DR - A vision document is not a tagline, a mission, or a roadmap. It’s a 2–5 year customer-centric story that recruits missionaries. According to SVPG’s Marty Cagan, most teams confuse vision with mission and lose the “common north star” benefit in the process. The 35 prompts below fix that, one section at a time.

Why your team is drifting in 2026 (the data is brutal)

Here’s the part I want you to read twice. The 2026 State of Product Management Report from ProductPlan surveyed 250 product leaders and surfaced something every founder should screenshot: over 60% of prioritization frameworks get overridden by leadership escalations (ProductPlan, 2026). Translation: even when a team agrees on a framework, executives blow it up because nobody can point back to a shared vision doc and say, “this bet contradicts the vision, kill it.”

Another stat from the same report: teams in the “transition stage” of product maturity perform worse on alignment than project-based teams. That’s a knife. The very act of trying to become more product-led, without a written vision, makes you more confused than the legacy team you were trying to replace.

Pull quote - “Over 60% of prioritization frameworks are overridden by leadership escalations.” - ProductPlan, State of Product Management Report 2026

If your team is shipping the wrong things faster than ever, the problem isn’t velocity. It’s vision. Let’s fix it.

The 7-part product vision doc anatomy (use this skeleton)

A vision document is a specific kind of artifact. Roman Pichler’s Product Vision Board gives us five clean sections (vision, target group, needs, product, business goals). Marty Cagan’s SVPG articles add a “visiontype” idea - the vivid experience that brings it to life. I blend both into a 7-part skeleton that fits a real early-stage company:

Vision headline - the 2-5 year “painted picture” of the customer’s new life
Mission - the company’s purpose, distinct from the vision
Target group & ICP - exactly who you serve, with attributes
Problem, value, and jobs-to-be-done - the specific pain you remove
Strategy & wedge - your unfair angle against alternatives
Roadmap, OKRs, and North Star - the outcomes you’ll measure
Communication, principles, and decision rules - how the doc gets used

Each prompt section below maps to one of these seven parts. Run them in order, paste output into Notion or Confluence, and you have a draft doc in two sittings.

Section 1 - Vision & mission prompts (prompts 1–5)

These five prompts write the top of the doc: the vision headline, the mission, and the customer “painted picture.” Marty Cagan warns that a vision is not a slogan like “organize the world’s information” - that’s a mission. Your vision has to describe how the customer’s life changes. The prompts below force that depth.

Prompt 1 - Vision headline generator (Marty Cagan style)

Purpose: Produce a 2–5 sentence vision that focuses on the customer’s new life, not on your product features.

You are Marty Cagan's product vision coach. I run a [STAGE] startup called
[COMPANY] in the [INDUSTRY] space. Our target customer is [ICP DESCRIPTION].
The problem we attack is [TOP 1-3 PROBLEMS]. The trend that finally makes this
solvable is [ENABLING TECH TREND, e.g. LLMs, vertical AI agents, etc.].

Write me a product vision that:
- Is 2-5 sentences long
- Describes the desired end state 2-5 years out
- Focuses on the customer's life, NOT on our product features
- Is specific enough to be falsifiable
- Is inspiring enough that a strong PM would quit their job to work on it
- Mentions the enabling technology trend that powers the shift

Return three variants: bold, balanced, and understated. For each, give me the
"visiontype" - a one-paragraph scene of a real user living the future state.

Example output (abbreviated):

Balanced variant: “In five years, independent marketing agencies will run client campaigns the way Netflix runs recommendations - by feeding the system ten years of first-party data, pressing a button, and shipping a fully-briefed creative package before their morning coffee ends. [COMPANY] makes that world real by turning every agency’s creative archive into a private generative engine that learns the agency’s taste, not the internet’s.”

Pro tips:

Reject any output that starts with “Our product is…” - that’s a feature, not a vision.
Paste the output into a doc, delete every adjective. What remains should still feel inspiring.
The “visiontype” paragraph is gold. Use it to brief designers for a Figma moodboard.

Prompt 2 - Mission vs vision separator

Purpose: Marty Cagan’s #1 complaint is that founders conflate mission and vision. Use this prompt to write both, side by side, so the team can finally tell them apart.

Here is my current one-paragraph "vision" statement: [PASTE].

You are an SVPG-trained product coach. Split it into TWO distinct statements:

1. MISSION: a 1-sentence company purpose (the "why we exist"). Timeless, no
   technology, no product mention. Format: "To [verb] [object] for [audience]
   so that [outcome]."

2. VISION: a 2-5 sentence customer-centric picture of the future state, 2-5
   years out, that is impossible without your specific approach.

For each, explain in 2 bullet points why the language is mission-grade or
vision-grade. Then flag any sentences from my original paragraph that should
be cut because they belong in a strategy doc, not a vision doc.

Example output: Your original “We help SMBs sell more using AI” splits into:

Mission: “To give every small business owner a revenue team they could never afford to hire.”
Vision: “In three years, a solo founder in Des Moines will out-sell a 20-person sales org in Austin - not by working harder, but by compounding 10 years of customer data into an AI co-seller that drafts, sends, and books meetings while the founder sleeps.”

Pro tips:

Stick both sentences on a Notion cover. Sales will use the mission, product will use the vision. That’s healthy.
If your “mission” can apply to 5 different companies, you still haven’t found it.

Prompt 3 - Vision critique (Cagan’s red-team)

Purpose: Marty Cagan is famous for saying “I have never been inspired by a canvas or board.” This prompt pressure-tests your draft.

Here is my draft product vision: [PASTE].
Here is my draft mission: [PASTE].
Our product does: [1 SENTENCE].

You are a senior product coach trained on Marty Cagan's SVPG articles and Roman
Pichler's vision board checklist. Score the vision on these 6 criteria using a
1-5 scale:

1. INSPIRING - would a strong PM quit their job to work on this?
2. SHARED - could it unite 5 different teams under one north star?
3. ENDURING - will it still be true 5-10 years from now?
4. CONCISE - can a new hire repeat it from memory after one read?
5. AMBITIOUS - does it describe a "big audacious goal" that may never be
   fully reached?
6. ETHICAL - does it produce no harm to people or planet?

For each criterion, give a 1-line score and a 1-line fix. Then return a final
"tough love" paragraph: what would Marty Cagan say to a founder who showed
him this draft at an offsite?

Example output: “Vision scores 4/5 on Inspiring, 2/5 on Shared (too feature-y), 1/5 on Enduring (mentions GPT-4 by name, which is dated). Marty’s tough love: ‘You’ve written me a roadmap slide. A roadmap is what we are doing this quarter. A vision is what makes an ordinary PM produce an extraordinary product. Try again.’”

Pro tips:

Run this prompt on at least 3 different models. The variance in scores is itself useful signal.
If you score under 4 on “Shared,” rewrite. A vision only matters if 50 people can describe it the same way.

Prompt 4 - Customer “painted picture” narrative

Purpose: The visiontype. The single most powerful vision artifact Marty Cagan describes. This prompt generates the 1-page customer story that makes the vision visceral.

You are a senior product designer and storyteller. Our vision is: [PASTE].
Our ICP is: [PASTE].
Our product is: [PASTE].

Write me a 1-page "Day in the Life" narrative of our ICP using our product
3 years from now. Requirements:
- Use a fictional but realistic name (e.g. "Maya, 34, head of growth at a
  40-person DTC brand in Austin")
- Set a specific day and time (e.g. "Tuesday, 7:14 AM, October 2029")
- Show 4 moments: the trigger, the workflow, the surprise, the outcome
- Include sensory detail (what does the screen look like, what does she
  hear, what does she text a colleague)
- End with the line that captures the entire vision in 1 sentence
- Avoid feature names; describe outcomes in human terms

Then return a "design brief" paragraph: 3 visual cues a designer should
explore for the visiontype video.

Example output: A 600-word narrative of Maya waking up, getting a Slack ping from her AI co-pilot, opening one tab, approving a campaign, sharing a Loom with her CEO, and closing her laptop at 4 PM. Ends with: “Maya didn’t sell more because she worked harder. She sold more because her AI co-seller did the parts of her job she used to hate.”

Pro tips:

Record a 60-second voice memo of you reading this aloud. That’s your “internal trailer.”
This narrative, not the bullet doc, is what you share with new hires in week one.

Prompt 5 - Vision-in-one-sentence stress test

Purpose: Forces extreme compression. If you can’t say it in 22 words, you don’t have it.

Here is my vision: [PASTE].
My ICP is: [PASTE].
The single biggest change in their life will be: [PASTE].

Compress my vision into a 22-word-or-fewer statement that survives a cocktail
party test. A stranger should be able to repeat it back to a friend without
notes. Then return 5 alternate 22-word versions, each with a different
emotional register: rebellious, calm, witty, urgent, poetic.

For each version, list the 2-3 words you would underline if you printed it on
a wall. Explain why those words earn the underlining.

Example output: “Calm: We make every small business a smart business.” “Rebellious: Kill the spreadsheet. Run your company like Netflix runs recommendations.” “Poetic: A revenue team in your pocket, no hiring required.”

Pro tips:

The cocktail party test is brutal. If your spouse can’t repeat it, keep cutting.
Don’t pick the witty one. Pick the one a customer would tattoo on a notebook. Boring clarity beats clever confusion.

Section 2 - Target & ICP prompts (prompts 6–10)

A vision without a target group is a vibe. Roman Pichler’s vision board checklist demands your target group be clear, specific, and cohesive - meaning you can tell who is in and who is out, and the people inside share enough attributes that you’d build the same product for all of them. These prompts turn fuzzy personas into a tight ICP.

Prompt 6 - ICP deep profile

Purpose: Build a one-page ICP that a sales rep could use to qualify a lead in 30 seconds.

You are a B2B positioning strategist trained on the April 2026
"product-strategy stack" model. Build a one-page ICP profile for our product.

Product: [PASTE]
Vision: [PASTE]
Beachhead market: [PASTE]
Price point: [PASTE]

For the ICP, return:
1. Firmographics (industry, size, revenue band, stage, geography)
2. Tech stack signals (tools they already use that we plug into)
3. Trigger events (the 3-5 things that have to be true in the next 90 days
   for them to start shopping)
4. Buyer roles (economic buyer, technical buyer, champion, blocker)
5. Anti-ICP (the 1-2 segments we WILL NOT serve, with one-sentence reasons)
6. A 3-line qualification snippet a SDR can memorize

Format as a table.

Example output: A table with rows for firmographics, tech stack, triggers, roles, anti-ICP, and the SDR snippet. The anti-ICP row is often the most valuable output - it tells your team what to ignore.

Pro tips:

Cut any ICP description that contains the word “anyone.” Your ICP is never everyone.
Run this twice: once for “who would pay us $1k/mo” and once for “who would pay us $50k/mo.” You’ll see different buyers.

Prompt 7 - Jobs-to-be-done interview script

Purpose: JTBD is the most overused phrase in product and the least used framework. This prompt generates a 30-minute interview script that actually surfaces jobs, not feature requests.

You are a JTBD expert trained on Bob Moesta and Tony Ulwick's frameworks.
Design a 30-minute customer interview script to uncover the "job" our ICP
hires our product to do.

Context:
- Product: [PASTE]
- ICP: [PASTE]
- We suspect the core job is: [HYPOTHESIS]

Generate 12 questions in this order:
- 2 warm-up questions
- 2 questions about the last time they tried to solve the problem WITHOUT us
- 3 questions about the trigger event that pushed them to look for a solution
- 2 questions about what "done" looks like (desired outcome)
- 2 questions about emotional and social dimensions
- 1 question about what they'd do if our category didn't exist

For each question, list:
- The question (open-ended, no leading language)
- The interviewer follow-up ("Tell me more about that…")
- The signal we are listening for
- The "JTBD statement" template we will fill in after the call:
  "When [situation], I want to [motivation], so I can [expected outcome]."

Return the script as a clean Notion-ready document.

Example output: A 12-question script. The interviewer fill-ins are what separate this from a generic survey. They force the human to keep going past the surface answer.

Pro tips:

Run 8 interviews in two weeks. That’s the minimum for a real signal.
After each call, paste the transcript into Prompt 8 below. Patterns appear fast.

Prompt 8 - Call transcript to JTBD extractor

Purpose: Turn raw customer call notes into JTBD statements. Save 10 hours of manual tagging.

You are a senior product researcher. I will paste raw customer call notes
below. Extract every JTBD statement using the format:
"When [situation], I want to [motivation], so I can [expected outcome]."

Also return:
1. Top 5 pain quotes (verbatim, with speaker pseudonym)
2. Top 3 desired outcome patterns
3. Top 3 unmet needs (frequency-ranked)
4. Top 2 unexpected jobs (ones we did not anticipate)
5. 3 follow-up questions to test in the next round of interviews

Be ruthless about quoting exact words. Do not paraphrase.

NOTES:
[PASTE 2,000-10,000 WORDS OF CALL NOTES]

Example output: Twelve JTBD statements, 5 pain quotes, 3 outcome patterns, 3 unmet needs, 2 surprise jobs, and 3 follow-up questions. The “unexpected jobs” row is the gold - it’s where differentiation lives.

Pro tips:

Always cross-check the “unexpected jobs” against your roadmap. If they show up in 3+ calls, they belong in a v2 spec.
Keep a running doc. After 8 calls, your JTBD list should be stable. If it isn’t, your ICP is wrong.

Prompt 9 - Persona-to-segment validator

Purpose: Roman Pichler warns about personas that aren’t cohesive. This prompt tests whether your “personas” are really one ICP or two.

You are a positioning expert trained on the April 2026 ProductPlan State of
PM Report. I have 4 personas in our product brief:

1. [PERSONA A - title, attributes, top job]
2. [PERSONA B - title, attributes, top job]
3. [PERSONA C - title, attributes, top job]
4. [PERSONA D - title, attributes, top job]

For each persona, return:
- A 1-sentence "are they the same person in different costumes?" verdict
- The 2 attributes that overlap most across personas
- The 1 attribute that splits them into a different ICP
- A recommendation: KEEP, MERGE with another persona, or DELETE

Then return a "rule of three" - the 3 personas that should remain in the
final doc, in priority order, with a one-sentence "if you only build for
one, build for…" statement for each.

Example output: “Persona C and D are the same buyer at different career stages - merge them. Persona A is actually an anti-ICP, delete.”

Pro tips:

Less is more. Three personas is plenty. Eight is a soap opera.
The “build for one” line is the most-quoted line in good vision docs.

Prompt 10 - Beachhead wedge finder

Purpose: Geoffrey Moore’s Crossing the Chasm idea - your beachhead has to be small enough to dominate. This prompt picks the wedge.

You are Geoffrey Moore's positioning coach in 2026. We have a vision and ICP
already. Now we need a beachhead.

Vision: [PASTE]
ICP: [PASTE]
Candidate beachheads we are considering:
- [BEACHHEAD 1]
- [BEACHHEAD 2]
- [BEACHHEAD 3]

For each, score 1-5 on:
1. Whole product - can we be the only vendor they need?
2. Competition - is the slot empty or full?
3. Reference-ability - will this buyer tell 5 peers about us?
4. Channel fit - can we reach them with our current GTM motion?
5. Expansion path - does winning here unlock 3 adjacent markets?

Return a ranked shortlist with the winning beachhead and a 90-day
domination plan: the 3 things we must own in this segment by day 90.

Example output: A scored table, the winning beachhead, and a 90-day plan like “be the only AI co-pilot in the ‘Series A fintech in Lagos’ segment by Q3 2026.”

Pro tips:

The winning beachhead is rarely the biggest market. It’s the one where you can be the only choice.
If all three score under 3 on Reference-ability, you’re picking the wrong wedge.

Section 3 - Problem & value prompts (prompts 11–15)

Now that you know who you’re serving, define the specific problem and the measurable value. The prompts in this section use StoryBrand’s “external problem vs internal problem vs philosophical problem” framing, plus Strategyzer’s Value Proposition Canvas. If you can’t describe the problem in one sentence that a stranger repeats back to you, the rest of the doc is fiction.

Prompt 11 - Three-layer problem statement (StoryBrand style)

Purpose: Customers buy solutions to three layers of problem. Most founders only address the first.

You are Donald Miller's StoryBrand coach. Write a three-layer problem
statement for our product.

Product: [PASTE]
ICP: [PASTE]
Surface complaint we hear in sales calls: [PASTE]

Return:
1. EXTERNAL problem - the practical, visible, day-to-day issue
   (e.g. "I waste 4 hours a week on manual reporting")
2. INTERNAL problem - the emotional, private frustration
   (e.g. "I feel like I'm the only one on the team who cares about this")
3. PHILOSOPHICAL problem - the wrong-in-the-world stakes
   (e.g. "Operators deserve to know if their work matters, not guess")

For each layer, give:
- A 1-sentence description
- A 1-sentence "we make this go away" promise
- A specific example a customer would text a friend

Then draft a 30-second elevator pitch that walks the three layers in order.

Example output: External / Internal / Philosophical, each with a promise and a friend-text example. The pitch reads as a story arc: surface pain → emotional toll → worldview stake.

Pro tips:

The philosophical problem is the one your investors and evangelists will quote. Don’t skip it.
Re-test all three with 3 customers in week 2. If they nod, ship it. If they shrug, rewrite.

Prompt 12 - Value Proposition Canvas (Strategyzer)

Purpose: Strategyzer’s Value Proposition Canvas is the cleanest way to map a job to a gain. This prompt fills it in.

You are Alex Osterwalder's Value Proposition Canvas coach. Build a
half-canvas (value map side) for our product.

Customer jobs (pasted from Prompt 7 outputs): [PASTE]
Pains (pasted from Prompt 8 outputs): [PASTE]
Gains (pasted from Prompt 8 outputs): [PASTE]

Return:
1. Products & services - our list, ranked by which pain they kill
2. Pain relievers - exact phrases that map each pain to a reliever
3. Gain creators - how each gain gets unlocked (e.g. "save 4 hours/week"
   becomes "auto-generate the Monday report by 6 AM")
4. A 1-paragraph "value statement" in this format:
   "We help [ICP] who want [job] by [product] that delivers [key gain]
   unlike [alternative]."

Then flag any gain we claim that we cannot prove with a screenshot or a
metric. Cut those claims.

Example output: A filled canvas, a clean value statement, and a list of unprovable claims to cut.

Pro tips:

The “cut unprovable claims” step is the most valuable. Marketing loves adjectives. Reality hates them.
Run this prompt in two modes: “aspirational” (what we want to be true) vs “today” (what we can ship). Reconcile the gap in your strategy doc.

Prompt 13 - 10x vs 10% value wedge

Purpose: Peter Thiel’s classic question - are you 10x better, or 10% better? This prompt forces the answer.

You are Peter Thiel's Blitzscaling coach. Describe the value wedge between
our product and the next best alternative for our ICP.

Alternative our ICP uses today: [PASTE]
Our product: [PASTE]
Key pain we kill: [PASTE]

Return:
1. A "10% better" version - small, incremental, defendable
2. A "10x better" version - bold, transformative, possibly delusional
3. A recommendation: which version is our real wedge in 2026, and why?
4. The single metric that proves the wedge (e.g. "time-to-first-campaign")
5. A 1-paragraph investor-friendly "why 10x" story

Be honest. If we are only 10% better, say so. Don't pretend.

Example output: “You are 10% faster on reporting and 10x better on creative generation. Lead with creative generation. The 10% story is for retention, the 10x story is for acquisition.”

Pro tips:

10x is hard. If you can’t credibly claim 10x on the outcome, claim 10x on a dimension (speed, cost, learning curve).
The “single metric” line is the seed of your North Star. Save it.

Prompt 14 - Customer pain → outcome ladder

Purpose: A table that maps every customer pain to a measurable business outcome. Sales loves this. Engineering tolerates it.

You are a product ops lead. Build a "pain → outcome" ladder for our product.

Pains (pasted): [PASTE]
Customer jobs (pasted): [PASTE]

Return a markdown table with columns:
- Pain (verbatim from customer calls)
- Frequency (1-5)
- Severity (1-5)
- Current workaround
- Time wasted per week (hours)
- Money wasted per year (USD)
- Our feature that kills the pain
- Outcome metric we will move

Sort by (Severity × Frequency). The top 3 rows are your Q1 roadmap.

Example output: A ranked table. The Q1 roadmap is the top 3 rows. The bottom rows become a “future bets” list.

Pro tips:

This table is the bridge between vision and roadmap. Show it to engineering, then ask: “Which of these would you be excited to ship?”
Re-rank every quarter. Pain shifts.

Prompt 15 - Anti-vision: what we will NOT do

Purpose: A vision doc is also defined by what it excludes. This prompt is the most underrated prompt in the entire library.

You are a strategy coach. Based on our vision and ICP, generate the
"anti-vision" - the 10 things our product will NEVER do, no matter how
many customers ask.

Vision: [PASTE]
ICP: [PASTE]
Common feature requests we keep getting: [PASTE]
Adjacent markets we are tempted by: [PASTE]

For each anti-vision item, return:
- The feature/market
- The honest reason we are saying no (e.g. "wrong ICP", "kills our
  wedge", "we'd have to rebuild the engine", "low margin")
- The single line we will tell a customer when they ask

Then return a 1-paragraph "principle" we will print on the wall that
captures the discipline.

Example output: A list of 10 “no’s,” each with a reason and a customer-facing line. The wall principle might read: “We say no to 9 out of 10 ideas. We say yes to the 1 that compounds our wedge.”

Pro tips:

Anti-vision is more important than vision. Saying no to 100 good ideas is what keeps a startup coherent.
Print the wall principle. Literally print it. Frame it. Read it in every planning meeting.

Section 4 - Strategy & wedge prompts (prompts 16–20)

A vision is a 2-5 year destination. Strategy is the next 12 months of moves that buy you a chance to reach it. These prompts combine Marty Cagan’s product strategy framing, Geoffrey Moore’s positioning, and the Strategyzer Business Model Canvas. The output here is what your board, leadership team, and head of product will all reference.

Prompt 16 - Strategy on a page

Purpose: Cagan says strategy must change at least yearly. This prompt produces a one-pager that fits on a slide.

You are a senior product strategy coach. Generate a "Strategy on a Page"
for the next 12 months, given:

Vision: [PASTE]
ICP: [PASTE]
Beachhead: [PASTE]
Anti-vision: [PASTE]
Top pains (from Prompt 14): [PASTE]

Return:
1. The 3 strategic bets (each with: hypothesis, owner, time horizon, kill
   criteria, success metric)
2. The 3 things we are explicitly NOT doing (linked to anti-vision)
3. The 2 big assumptions we are validating this quarter
4. The 1 metric the entire company will be measured on
5. The 1 sentence a new hire should be able to repeat about our strategy

Format as a single Notion page with H2 headers.

Example output: Three bets, three no-go’s, two assumptions, one metric, one sentence. The one-sentence line is the strategy equivalent of a vision headline.

Pro tips:

“Kill criteria” is the most important column. If a bet has no kill criteria, you have a religion, not a strategy.
Re-run this prompt every quarter. Strategy should evolve; vision shouldn’t.

Prompt 17 - Business Model Canvas (Strategyzer)

Purpose: Validate the unit economics before you scale.

You are Alex Osterwalder. Build a Business Model Canvas for our startup.

Vision: [PASTE]
ICP: [PASTE]
Beachhead: [PASTE]
Pricing hypothesis: [PASTE]
Top channels: [PASTE]

Return the 9 blocks: customer segments, value propositions, channels,
customer relationships, revenue streams, key resources, key activities,
key partnerships, cost structure.

For each block:
- 3-5 bullet points
- The single most uncertain assumption in that block
- The cheapest test we can run in 2 weeks to validate that assumption

Highlight the 2 blocks with the highest cumulative uncertainty. Those
are the blocks that need founder attention this month.

Example output: A filled canvas, with 9 “single most uncertain assumption” bullets and 9 two-week tests. The 2 highest-uncertainty blocks become the focus.

Pro tips:

The cheapest test is rarely a survey. It’s a 5-customer phone call or a $200 smoke-test landing page.
Revisit the canvas every 6 months. Pricing almost always changes first.

Prompt 18 - Competitive wedge map

Purpose: Build a 2x2 map that shows why you win where you win.

You are a competitive strategy coach. Map our wedge against the top 4
alternatives our ICP considers.

Our product: [PASTE]
Top 4 competitors/alternatives: [PASTE]
ICP: [PASTE]

Return a 2x2 map with these axes:
- X axis: "Implementation effort" (low → high)
- Y axis: "Outcome specificity" (generic → ICP-specific)

Plot all 5 options. Then for each of the 4 competitors, give:
- Their stated positioning in 1 sentence
- Their biggest blind spot
- The 1 feature/marketing line that exposes it

Then recommend the single wedge we should defend and the single wedge
we should attack.

Example output: A markdown table simulating a 2x2 plot, the 4 competitor cards, and the defend/attack recommendation.

Pro tips:

The “biggest blind spot” is your wedge. Repeat it on every landing page.
Update this map twice a year. New entrants appear faster than you’d think.

Prompt 19 - Pricing-strategy alignment check

Purpose: Pricing is a vision artifact, not a finance artifact. Make sure yours aligns.

You are a pricing strategist trained on the 2026 OpenView and ProductPlan
benchmarks. Audit our pricing against our vision and ICP.

Vision: [PASTE]
ICP: [PASTE]
Pricing: [PASTE]
Anti-vision: [PASTE]
Customer pains (from Prompt 14): [PASTE]

Return:
1. The 3 messages our pricing sends to the ICP (intended or not)
2. The 1 message that contradicts our vision
3. A pricing experiment we can run in 30 days to test the wedge
4. The 2 features we should NEVER charge extra for (because they are
   table-stakes for our vision)
5. The 1 feature we should consider charging 2-3x more for (because it
   disproportionately delivers the vision)

Example output: A pricing audit, a 30-day experiment design, and 2/1 lists of “free this, premium that.”

Pro tips:

If your pricing page lists 4 tiers, you have a roadmap problem, not a pricing problem. Cut to 2.
The “premium for the dream outcome” insight is the most-leveraged pricing move in SaaS right now.

Prompt 20 - Risks and assumptions log (Cagan’s 4 risks)

Purpose: Marty Cagan teaches that every product has 4 risks: value, usability, feasibility, viability. This prompt makes them explicit.

You are Marty Cagan. List the 4 product risks for our biggest bet this
quarter.

Vision: [PASTE]
Strategic bet: [PASTE]
Team: [PASTE]
Tech stack: [PASTE]

Return a table with columns: Risk Type, Specific Risk, Severity (1-5),
Cheapest Test, Owner, Date of First Test.

Risk types: VALUE (will they buy/use it?), USABILITY (can they figure it
out?), FEASIBILITY (can we build it?), VIABILITY (does it work for the
business?)

For each risk, the cheapest test should take 1-5 days, cost under $500,
and produce a yes/no signal.

Example output: A risks table, with each risk mapped to a 1-week test. The table is the agenda for the next sprint.

Pro tips:

If you don’t have a viability risk, your bet isn’t bold enough.
Review the table every Friday. The cheapest test that wasn’t run is your biggest opportunity.

Section 5 - Roadmap & OKR prompts (prompts 21–25)

Now translate the strategy into a roadmap and OKRs. The 2026 ProductPlan report flags that “becoming more outcome-focused” is one of the top challenges for product teams. These prompts force the shift from output to outcome. The output here is what your eng team, board deck, and quarterly review will all use.

Prompt 21 - Quarterly OKR builder

Purpose: Three objectives, each with three key results. The classic Google structure, updated for 2026.

You are a Google-trained OKR coach. Build Q[X] OKRs for our team.

Vision: [PASTE]
Strategy on a page: [PASTE from Prompt 16]
Top pains (from Prompt 14): [PASTE]
Available team: [X engineers, Y PMs, Z designers]

Generate 3 objectives. For each objective:
- 1-sentence objective (qualitative, time-bound, inspiring)
- 3 key results (quantitative, measurable, ambitious but believable)
- The "if we hit this, what does it unlock next quarter?" line
- The 1 metric we are willing to be publicly graded on

Then return the OKRs in a markdown table, sorted by impact. Flag any KR
that is actually a deliverable disguised as an outcome. Replace those
with outcome-based KRs.

Example output: Three objectives, nine KRs, three “unlocks” lines. The flagged deliverables (e.g. “ship feature X”) get rewritten as outcomes (e.g. “increase activation rate from 22% to 35%”).

Pro tips:

A KR that ends in ”%” or ”$” or a number is almost always an outcome. A KR that ends in “shipped” is a feature.
Don’t have more than 3 objectives per team. If you do, you have a focus problem.

Prompt 22 - Now/Next/Later roadmap

Purpose: The most-loved lightweight roadmap format. Works in Notion, Tana, Linear, and Slite.

You are a product ops coach. Turn our top 3 strategic bets and our top
pains (from Prompt 14) into a Now/Next/Later roadmap.

Strategic bets: [PASTE]
Top pains: [PASTE]
Current capacity: [X sprints of work]

Return:
- NOW: 3-5 items shipping in the next 6 weeks
  (each with: name, problem solved, success metric, owner)
- NEXT: 3-5 items in the next 6-12 weeks
  (each with: hypothesis we are validating, kill criteria)
- LATER: 3-5 bets for the next 12-24 weeks
  (each with: why it is later, the assumption we need to validate first)

Then return a 1-paragraph "what's NOT on this roadmap and why" - the
explicit trade-offs.

Example output: A clean Now/Next/Later table, with a trade-off paragraph that defends the choices.

Pro tips:

The trade-off paragraph is what your CEO will quote in the all-hands. Write it carefully.
Update Now/Next/Later every 2 weeks. Now becomes Done, Next becomes Now, Later moves up.

Prompt 23 - Outcome-based PRD header

Purpose: Stop writing PRDs that read like feature specs. Use this prompt for the first half-page.

You are a product manager trained on the Continuous Discovery model
(Teresa Torres, 2026). Write the outcome-focused header for a PRD for
[FEATURE].

Vision: [PASTE]
ICP: [PASTE]
Pain we are killing: [PASTE]
Outbound hypothesis (from discovery): [PASTE]
Discovery evidence: [PASTE]

Return a PRD header with:
1. Problem (1 paragraph, customer-quote-led)
2. Outcome (the metric we will move, baseline → target)
3. Non-goals (3 bullets - what this PRD is NOT trying to solve)
4. Risks (value / usability / feasibility / viability - 1 line each)
5. Discovery artifacts (the tests we ran and what we learned)
6. Success criteria (the date and the number that will tell us we won)

Example output: A clean PRD header that reads like a strategy doc, not a Jira ticket. The non-goals section is the most valuable part.

Pro tips:

Paste this into Linear, Notion, or your PRD tool. It will save your eng team 5 questions per ticket.
A PRD with no “non-goals” is a feature request, not a product decision.

Prompt 24 - Discovery plan (Teresa Torres continuous discovery)

Purpose: Teresa Torres’ continuous discovery asks for one customer interview per week per PM. This prompt produces the weekly cadence.

You are Teresa Torres. Design a continuous discovery plan for our team.

Vision: [PASTE]
ICP: [PASTE]
Current team: [X PMs, Y designers, Z engineers]
Top opportunities: [PASTE]
Discovery cadence target: 1 interview per PM per week

Return:
1. The opportunity solution tree (1 root, 3 opportunities, 3 solutions
   per opportunity, 3 tests per solution - visually described)
2. The weekly interview template (15-min, structured)
3. The "insights board" format (where we capture, who reviews, when we
   synthesize)
4. The 4-week cadence: week 1 interviews, week 2 synthesis, week 3
   prototype, week 4 test
5. The 1 question the team asks in every retro: "What did we learn this
   week that we didn't know last week?"

Format as a single Notion page.

Example output: A continuous discovery plan, with the cadence built in. The retro question is the most important line - it’s how culture shifts from output to learning.

Pro tips:

Track interview count per PM. If a PM is below 3/month, they’re not doing discovery.
The opportunity solution tree is a living artifact. Pin it in the team channel.

Prompt 25 - Quarterly business review (QBR) builder

Purpose: A QBR is just the vision + strategy + OKRs revisited with data. This prompt drafts the deck.

You are a chief of staff to a startup CEO. Draft the structure and key
slides for our Q[X] QBR.

Vision: [PASTE]
Strategy on a page: [PASTE from Prompt 16]
OKRs: [PASTE from Prompt 21]
North Star and supporting metrics: [PASTE]
Customer wins and losses from this quarter: [PASTE]

Return a slide-by-slide outline (10-12 slides):
- Cover (quarter, the 1-line vision, the 1-line strategy)
- North Star trend (3-quarter chart, 1-line interpretation)
- OKR scorecard (3 objectives, 3 KRs each, RAG status)
- Top 3 wins (with customer quote, metric, dollar impact)
- Top 3 misses (with root cause, what we learned, fix for next quarter)
- Customer signal (top 3 pains, top 3 gains, top 3 churn reasons)
- Roadmap deltas (what we added, cut, deferred)
- Strategic bet review (3 bets, status, kill/continue decision)
- Next quarter OKRs (preview)
- Asks (from the board: hires, capital, intros)
- Appendix (granular metrics, cohort analysis)

For each slide, return the headline bullet and the supporting data needed.

Example output: A full slide outline with the data inputs. The “misses” slide is the most important - leadership teams that own their misses earn trust.

Pro tips:

The “asks” slide is what most boards remember. Be specific. “Intros to 3 fintech CFOs in NYC” beats “help with intros.”
Send the deck 24 hours before the QBR. Real time should be discussion, not reading.

Section 6 - Metrics & North Star prompts (prompts 26–30)

Without a North Star Metric, your team optimizes the wrong things. The 2026 ProductPlan report confirms what Lenny Rachitsky has been saying for years - companies with a clear North Star grow faster. These prompts pick yours, define the supporting metric tree, and wire it into dashboards in Notion, Productboard, or Aha!.

Prompt 26 - North Star Metric selector

Purpose: The North Star Metric is the single number that, if it goes up, means customers are getting value AND the business is healthy.

You are Lenny Rachitsky and the Amplify Partners team. Help us pick our
North Star Metric (NSM).

Vision: [PASTE]
ICP: [PASTE]
Business model: [PASTE - SaaS, marketplace, fintech, etc.]
Current funnel: [PASTE]
What we are optimizing for in 2026: [PASTE]

Return:
1. The 3 candidate NSMs, each with a 1-sentence "why this" argument
2. For each, the supporting input metrics (3-5) that move the NSM
3. For each, the 2 "false growth" risks (e.g. NSM goes up but business
   is sick)
4. The 1 recommended NSM with a 3-paragraph justification
5. The 1 anti-pattern: the metric we will NOT pick, and why

Then describe the daily/weekly/monthly cadence: who checks the NSM when.

Example output: Three candidates, with risk profiles, and a clear recommendation. The “false growth” risks are the most underrated output.

Pro tips:

A good NSM is a value metric, not a revenue metric. Revenue is the lagging indicator. Value is the leading one.
If your NSM is the same as your revenue, you don’t have a North Star. You have a finance metric.

Prompt 27 - Input metric tree

Purpose: A North Star is useless without the input metrics that move it. This prompt builds the tree.

You are a growth analytics lead. Build the input metric tree for our
North Star.

North Star: [PASTE]
Funnel: [PASTE]
Top growth levers: [PASTE]

Return a tree (markdown indent) with:
- Level 1: North Star
- Level 2: 3-5 primary drivers
- Level 3: 2-3 sub-metrics per driver
- Level 4: the experiments we will run on each sub-metric this quarter

For each metric, include:
- Definition (numerator / denominator, time window)
- Current baseline
- 2026 target
- Owner

Then return a 1-paragraph "metric hygiene" note: which metrics are
suspect, which are double-counted, and which are vanity.

Example output: A 4-level tree, every metric with a baseline, target, and owner. The “metric hygiene” paragraph forces honesty.

Pro tips:

An owner without authority to change the metric is a vanity metric. Cut it.
The 2026 target should scare you a little. If it doesn’t, it’s too low.

Prompt 28 - Cohort and retention analyzer

Purpose: Cohorts tell you if the product is actually getting better. This prompt interprets your retention data like a senior analyst.

You are a retention analyst. Interpret the following cohort data and
surface the 3 most important findings.

Cohort table:
[PASTE monthly cohort retention, e.g. signup month × activity month]

Vision: [PASTE]
Product: [PASTE]
NSM: [PASTE]

Return:
1. The 1-sentence headline (good news or bad news, no hedging)
2. The 3 cohorts with the best and worst retention, and why
3. The inflection month where retention stabilizes (or doesn't)
4. The 2 product changes that likely caused the biggest retention deltas
5. The 1 experiment we should run next quarter to lift retention 10%

If the data is too thin, tell me what data to pull instead.

Example output: A clean retention analysis with actionable next steps. The “inflection month” insight is the one that changes roadmap priorities.

Pro tips:

Track retention by acquisition channel, not just by signup month. Channel quality drives everything.
If you can’t see an inflection by month 3, your product may have a habit problem. That’s a vision issue, not a metrics issue.

Prompt 29 - Quality and counter-metric guardrails

Purpose: Lenny warns that a North Star without counter-metrics is a speedometer without a brake. This prompt adds the brakes.

You are a quality-focused product leader. Define the 5 counter-metrics
that prevent our North Star from being gamed.

North Star: [PASTE]
Current quality signals: [PASTE]
Known failure modes: [PASTE]

For each counter-metric, return:
- Name (e.g. "Refund rate per cohort")
- Definition
- Threshold (green / yellow / red)
- The "if this goes red, the action we take" rule
- The "false alarm" pattern (when red is actually fine)

Then return a 1-paragraph "guardrail review" cadence: who reviews
counter-metrics weekly, and what the escalation path is.

Example output: Five counter-metrics with thresholds, action rules, and a cadence. The escalation path is what turns metrics from dashboards into decisions.

Pro tips:

Counter-metrics are not “negative metrics.” They are quality indicators that protect the North Star.
If you can’t write the action rule, you don’t understand the metric. Cut it.

Prompt 30 - Executive dashboard wireframe

Purpose: The 1-page dashboard your CEO checks every Monday. Build it once, refine forever.

You are a product ops lead. Design a 1-page executive dashboard for our
NSM and supporting metrics.

NSM: [PASTE]
Input metrics: [PASTE from Prompt 27]
Counter-metrics: [PASTE from Prompt 29]
Top 3 strategic bets: [PASTE]
Current quarter OKRs: [PASTE]

Return a wireframe with these zones:
- Top: NSM trend, 3-month sparkline, % change WoW
- Top right: 3 counter-metric sparklines (green/yellow/red)
- Middle: 3 strategic bet scorecards (RAG)
- Middle right: 3 OKR progress bars
- Bottom: Top 5 customer signal bullets (from interviews, support, sales)
- Bottom right: Asks and risks (1-2 lines each)

For each zone, list the data source, the refresh frequency, and the
owner. Keep the whole thing under 30 seconds to read.

Example output: A wireframe with data sources and owners. The “30-second rule” is what separates a dashboard from a data dump.

Pro tips:

If the dashboard takes more than 30 seconds to read, it’s a report, not a dashboard. Cut.
Build this in Notion, Productboard, Aha!, or Slite. Don’t waste eng time on a custom solution in the first 12 months.

Section 7 - Communication & alignment prompts (prompts 31–35)

The last 20% of a vision doc is communication. A vision nobody reads is a wish. These prompts produce the all-hands narrative, the team rituals, the recruiting pitch, and the investor one-liner. The 2026 ProductPlan report flags that “a confidence gap is preventing product strategy from landing with Sales and Customer Success” - these prompts are how you close that gap.

Prompt 31 - All-hands narrative

Purpose: A 12-minute talk the CEO can give every quarter to re-align 50 people. Marty Cagan calls this the “evangelism” job of the product leader.

You are a startup CEO. Write a 12-minute all-hands talk that re-aligns
the team on our vision, strategy, and current quarter.

Vision: [PASTE]
Strategy on a page: [PASTE]
Q[X] OKRs: [PASTE]
Top 3 customer wins: [PASTE]
Top 1 customer pain we are still failing on: [PASTE]

Structure the talk as:
- 1-minute cold open: a customer quote that captures the vision
- 2-minute vision reminder (the "painted picture" from Prompt 4)
- 2-minute strategy reminder: the 3 bets and why
- 2-minute OKR scorecard
- 2-minute customer signal: the wins and the pain
- 1-minute ask: the 1 thing every team member should do this week
- 1-minute close: a line that ends on the vision

Write it as a script, not a deck. Return slide suggestions in
parentheses. Total length: ~1,800 words.

Example output: A real script the CEO can read. The cold-open quote is the part the team will remember on Friday.

Pro tips:

Record the talk on video. New hires watch it in week one.
Update the customer quote every quarter. Same vision, new proof.

Purpose: The 2026 ProductPlan report found a confidence gap between product strategy and Sales. This prompt closes it.

You are a sales enablement leader. Build a 1-page sales one-pager that
turns our vision and strategy into customer-facing language.

Vision: [PASTE]
ICP: [PASTE]
Top pains: [PASTE]
Value wedge: [PASTE from Prompt 13]
Common objections: [PASTE]

Return:
1. The 3-sentence elevator pitch (sales-memo-friendly)
2. The 5 discovery questions (open-ended, JTBD-flavored)
3. The 5 objection handlers (objection → acknowledgment → reframe →
   proof)
4. The 2 case-study snippets (1 line each, with a hard metric)
5. The 1 disqualifier: "we are NOT a fit if…"
6. The "always" and "never" rules (2 of each)

Format as a single A4 page. Tight, no fluff.

Example output: A 1-page sales sheet that fits in a back pocket. The disqualifier is the most valuable line - it stops bad deals before they start.

Pro tips:

Refresh the one-pager every quarter, even if nothing changed. The act of refreshing is the alignment ritual.
Track which objections show up most. If one shows up 5+ times, your product has a real gap, not a sales gap.

Prompt 33 - Recruiting pitch and JD

Purpose: Marty Cagan says vision is your #1 recruiting tool. A vision doc that doesn’t show up in your job descriptions is a wasted doc.

You are a startup founder hiring for a [ROLE]. Write the recruiting
pitch section of the JD using our vision and strategy.

Vision: [PASTE]
Strategy: [PASTE]
Top 3 things this person will own: [PASTE]
The team they will join: [PASTE]

Return:
1. The 3-sentence "why this role exists" pitch
2. The 3 reasons a great person would want this job
3. The 1 reason a great person would NOT want this job (honesty builds
   trust)
4. The 5 outcomes this person will be measured on in year 1
5. The 1 line we will say in the final interview that makes the
   candidate say yes

Tone: warm, specific, no corporate fluff. Read it aloud. If it sounds
like LinkedIn, rewrite.

Example output: A JD intro that sounds like a human wrote it. The “why you would NOT want this job” line is radical - and it filters better than any ATS.

Pro tips:

A candidate who says yes to a vague pitch quits in 6 months. A candidate who says yes to a specific pitch stays 4 years.
Send the vision doc with the JD. If they don’t read it, they aren’t your person.

Prompt 34 - Investor one-liner and 5-sentence narrative

Purpose: YC, angels, and Series A partners hear 1,000 pitches a year. The first 30 seconds are everything.

You are a YC partner in 2026. Help us write a tight investor narrative
based on our vision doc.

Vision: [PASTE]
ICP: [PASTE]
Beachhead: [PASTE]
NSM and current value: [PASTE]
Stage and ask: [PASTE]

Return:
1. The 1-sentence pitch (under 30 words, no jargon)
2. The 5-sentence narrative (problem, insight, solution, traction, ask)
3. The 3 "founder advantage" bullets (why us, why now)
4. The 2 metrics we will lead with in the deck
5. The 1 question an investor will ask that we are NOT prepared for,
   and a 3-sentence answer

Example output: A 30-word one-liner, a 5-sentence narrative, and a “tough question” answer. The tough-question answer is the most important - it shows the founder has done the work.

Pro tips:

The 30-word pitch is what goes in the subject line of your cold email. Test it on 5 strangers. If they can’t repeat it, rewrite.
The “tough question” is the one that keeps you up at night. If you can answer it in 3 sentences, raise.

Prompt 35 - Decision-rule generator

Purpose: A vision doc is most useful when it’s used. Decision rules turn abstract vision into concrete yes/no answers in 50-person meetings.

You are a product ops coach. Turn our vision and strategy into 10
"decision rules" - the yes/no rules we will use in every planning
meeting.

Vision: [PASTE]
Strategy: [PASTE]
Anti-vision: [PASTE from Prompt 15]
NSM: [PASTE]

Generate 10 decision rules in this format:
"[TRIGGER] → if [test] then [action], because [principle]."

Examples of the shape:
- "New feature request → if it does not move the NSM, then park it,
  because we cannot afford to optimize two things at once."
- "New market request → if it requires rebuilding the engine, then
  decline, because the wedge is more important than the TAM."

Make 3 of the 10 rules "always" rules (do this no matter what).
Make 3 of the 10 rules "never" rules (do not do this no matter what).
Make 4 of the 10 rules "kill" rules (specific triggers that mean we
abandon an initiative).

Print these on a single page. Tape to the wall.

Example output: Ten rules, each with a trigger, test, action, and principle. The wall becomes the doc. The doc becomes the behavior.

Pro tips:

The “kill” rules are the most important. Without them, zombies live forever.
Re-derive the rules every 6 months. Vision doesn’t change, but the operationalization does.

Comparison table: 35 prompts at a glance

A picture is worth 35 prompts. Here’s the full library mapped to doc section, output format, and the primary framework each one draws from.

#	Prompt	Doc section	Output format	Primary framework	Approx. effort
1	Vision headline generator	Vision	3 variants + visiontype	Marty Cagan, SVPG	30 min
2	Mission vs vision separator	Mission	2 statements + rationale	Cagan, SVPG	20 min
3	Vision critique	Vision	6-criterion scorecard	Cagan, Pichler	25 min
4	Customer painted picture	Vision	600-word narrative	Cagan visiontype	45 min
5	22-word stress test	Vision	5 alt versions	Cocktail party test	20 min
6	ICP deep profile	Target group	One-page table	ProductPlan, April 2026	30 min
7	JTBD interview script	Customer jobs	12-question script	Moesta, Ulwick	45 min
8	Call transcript → JTBD	Customer jobs	12 statements + patterns	Continuous discovery	30 min
9	Persona validator	Target group	Merge/keep/delete list	Pichler checklist	25 min
10	Beachhead wedge finder	Beachhead	Scored table + 90-day plan	Geoffrey Moore	30 min
11	Three-layer problem	Problem	External/internal/philo	StoryBrand	25 min
12	Value Proposition Canvas	Value	Half-canvas	Strategyzer	40 min
13	10x vs 10% wedge	Wedge	Comparison + recommendation	Thiel, Blitzscaling	25 min
14	Pain → outcome ladder	Problem	Ranked table	Product ops	35 min
15	Anti-vision	Strategy	10 “no’s” + wall principle	Cagan, Moore	30 min
16	Strategy on a page	Strategy	Notion page	Cagan, SVPG	45 min
17	Business Model Canvas	Business model	9 blocks + tests	Strategyzer	60 min
18	Competitive wedge map	Wedge	2x2 + 4 cards	Competitive strategy	35 min
19	Pricing-strategy audit	Business model	Audit + 30-day test	OpenView, ProductPlan	30 min
20	4-risks log	Strategy	RAG table	Cagan, SVPG	25 min
21	Quarterly OKR builder	OKRs	3 objectives + 9 KRs	Google OKRs	40 min
22	Now/Next/Later roadmap	Roadmap	3-column table	Lightweight roadmap	35 min
23	Outcome-based PRD header	Roadmap	6-section header	Continuous discovery	25 min
24	Discovery plan	Discovery	OST + cadence	Teresa Torres	60 min
25	QBR deck	OKRs	12-slide outline	Board reporting	60 min
26	North Star selector	Metrics	3 candidates + rec	Lenny, Amplify	35 min
27	Input metric tree	Metrics	4-level tree	Growth analytics	40 min
28	Cohort analyzer	Metrics	5 findings	Retention analysis	30 min
29	Counter-metrics	Metrics	5 guardrails	Quality metrics	30 min
30	Executive dashboard	Metrics	Wireframe	Product ops	30 min
31	All-hands narrative	Communication	1,800-word script	Cagan evangelism	60 min
32	Sales one-pager	Communication	A4 page	Sales enablement	40 min
33	Recruiting pitch	Communication	JD intro	Talent	30 min
34	Investor narrative	Communication	30-word + 5-sentence	YC	30 min
35	Decision rules	Communication	10 rules	Operationalization	30 min

Total elapsed time if you run all 35 in sequence: roughly 17 hours. Spread across 14 days, that’s about 75 minutes a day - doable for one founder, even better with a co-founder pair.

14-day “vision reset” workshop agenda

Run this with your co-founder, head of product, and a designer. Time blocks are 90-minute sessions. Total: ~17 hours of focused work, plus async iteration.

Day 1 (Mon) - Vision headline & mission. Run Prompts 1, 2, 3. End with a 22-word vision (Prompt 5) and a 1-sentence mission. Write both on the wall.

Day 2 (Tue) - Painted picture. Run Prompt 4. The founder records a 60-second voice memo of the visiontype. The designer starts a Figma moodboard.

Day 3 (Wed) - ICP and beachhead. Run Prompts 6, 9, 10. End with a 1-page ICP and a single beachhead. Add the anti-ICP to a Notion page.

Day 4 (Thu) - Customer calls. Run 3 customer interviews using Prompt 7. Async-transcribe with a tool like Otter or Granola. Don’t prompt anything yet - just listen.

Day 5 (Fri) - JTBD synthesis. Run Prompt 8 on the transcripts. End with a list of 10 JTBD statements, ranked by frequency. Share in #product Slack.

Day 6 (Sat, async) - Reflection. Founder reads all outputs, hand-writes the “why this matters to me” memo. 30 minutes. Don’t prompt.

Day 7 (Sun) - Rest. No prompts. Vision is a creative act, not a grind.

Day 8 (Mon) - Problem & value. Run Prompts 11, 12, 13, 14. End with a Value Proposition Canvas, a 10x wedge statement, and a pain-to-outcome ladder.

Day 9 (Tue) - Anti-vision. Run Prompt 15. The “no’s” go on the wall next to the vision. Tension is healthy.

Day 10 (Wed) - Strategy & business model. Run Prompts 16, 17, 18, 19, 20. End with a Strategy on a Page, a Business Model Canvas, a competitive map, and a 4-risks log.

Day 11 (Thu) - Roadmap & OKRs. Run Prompts 21, 22, 23, 24. End with Q1 OKRs, a Now/Next/Later, a PRD template, and a discovery cadence.

Day 12 (Fri) - Metrics & dashboard. Run Prompts 26, 27, 28, 29, 30. End with a North Star Metric, a metric tree, a counter-metric list, and a dashboard wireframe.

Day 13 (Sat) - Communication. Run Prompts 31, 32, 33, 34, 35. End with an all-hands script, a sales one-pager, a recruiting pitch, an investor narrative, and 10 decision rules.

Day 14 (Sun) - Review & ship. Founder reads the entire doc out loud. Rewrite any sentence that doesn’t survive the read-aloud test. Print it. Bind it. Distribute it. Update the Notion, Confluence, or Slite master.

The output is a 30-page vision document, a 1-page summary, a dashboard, a sales sheet, a recruiting pitch, and an investor narrative. That’s the operating system for the next 12 months.

Common mistakes to avoid (the ones I’ve seen kill vision docs)

Confusing vision with mission. “Organize the world’s information” is a mission. Cagan’s vision-vs-mission piece is mandatory reading.
Writing a feature list. “An AI-powered CRM that…” is a product description, not a vision. Vision = customer’s new life.
Letting every team have their own “north star.” This breaks the alignment benefit. One vision, one company, every team subscribes to it.
Confusing the vision with a roadmap. Vision is 2-5 years. Roadmap is 1-2 quarters. The vision is the why of the roadmap.
Giving up at 6 months. Cagan warns that most startups fail not because their vision was bad, but because they gave up on it too early.
Skipping the visiontype. A vision without a vivid customer story is a slide. A vision with a visiontype is a movement.
Writing it alone. The founder drafts; the leadership team stress-tests. No exceptions.
Forgetting the anti-vision. A vision without explicit “no’s” gets diluted to nothing by month 6.
No rituals. A doc that lives in Notion is dead on arrival. The vision must be quoted, posted, and referenced.
No decision rules. A vision without a “if X, then Y” rule is a poster. With the rule, it’s a tool.

If you catch yourself doing any of these, stop. Go back to the relevant prompt. Rewrite.

Final word: your vision is a bet, not a slogan

Here’s the uncomfortable truth. Your vision is a bet. You’re betting the next 3-5 years of your life, your team’s energy, and your investors’ capital on a picture of the future that might happen. The vision document is the artifact that lets other people bet on you.

The 35 prompts above are the structure. They’re not the soul. The soul is the moment in Prompt 4 where you write down what your customer’s life looks like in 2029, and you read it back to yourself, and you feel something in your chest. That’s the moment you know you have a vision worth writing down.

Run the prompts. Iterate. Argue with the model. Throw out the bad drafts. Keep going until the document makes a strong PM quit their job to join you. That’s the test. That’s the only test.

If you want a starting point, copy Prompt 1 into ChatGPT, fill in the brackets, and run it before you close this tab. Twenty minutes from now, you’ll either have a vision or you’ll know what questions to ask. Either way, you’re further along than you were ten minutes ago.

Now go build the future your customers are waiting for.

Reader disclosure & educational-purpose notice

This page is published by SuperFreshAI for general informational and educational purposes only. By reading it, you agree to the points below.

Editorial independence. All reviews, guides, and recommendations are written by our editorial team based on hands-on use. Some links on this site are affiliate links, and some articles are produced as partner content — both are always clearly labeled. Our editorial conclusions are never shaped by partners or affiliates.
Not professional advice. Nothing on this page constitutes legal, financial, medical, tax, or other professional advice. AI tools, pricing, and capabilities change quickly — always verify current information with the tool's official documentation before making a decision.
Educational purpose only. The content here is intended to help you learn about AI tools and workflows. It is not a guarantee of results, performance, fitness for a particular purpose, or suitability for your specific situation. Your results may vary.
No warranties. The site and its content are provided on an "as is" and "as available" basis. We make no warranties, express or implied, about accuracy, completeness, reliability, or availability. See our Terms and Privacy for the full legal terms.
Your responsibility. You are responsible for how you use the information on this page, including any decisions you make based on it. Always do your own research and consult a qualified professional when appropriate.
Affiliate & partner disclosure. When you click certain outbound links, we may earn a commission at no extra cost to you. When a piece of content is produced as partner content, it is labeled at the top of the page. See our Editorial Policy for the full standards we follow.

By continuing to read, you acknowledge that you have read and understood this notice.

35 ChatGPT prompts for startup founders to write a clear product vision document

Why your team is drifting in 2026 (the data is brutal)

The 7-part product vision doc anatomy (use this skeleton)

Section 1 - Vision & mission prompts (prompts 1–5)

Prompt 1 - Vision headline generator (Marty Cagan style)

Prompt 2 - Mission vs vision separator

Prompt 3 - Vision critique (Cagan’s red-team)

Prompt 4 - Customer “painted picture” narrative

Prompt 5 - Vision-in-one-sentence stress test

Section 2 - Target & ICP prompts (prompts 6–10)

Prompt 6 - ICP deep profile

Prompt 7 - Jobs-to-be-done interview script

Prompt 8 - Call transcript to JTBD extractor

Prompt 9 - Persona-to-segment validator

Prompt 10 - Beachhead wedge finder

Section 3 - Problem & value prompts (prompts 11–15)

Prompt 11 - Three-layer problem statement (StoryBrand style)

Prompt 12 - Value Proposition Canvas (Strategyzer)

Prompt 13 - 10x vs 10% value wedge

Prompt 14 - Customer pain → outcome ladder

Prompt 15 - Anti-vision: what we will NOT do

Section 4 - Strategy & wedge prompts (prompts 16–20)

Prompt 16 - Strategy on a page

Prompt 17 - Business Model Canvas (Strategyzer)

Prompt 18 - Competitive wedge map

Prompt 19 - Pricing-strategy alignment check

Prompt 20 - Risks and assumptions log (Cagan’s 4 risks)

Section 5 - Roadmap & OKR prompts (prompts 21–25)

Prompt 21 - Quarterly OKR builder

Prompt 22 - Now/Next/Later roadmap

Prompt 23 - Outcome-based PRD header

Prompt 24 - Discovery plan (Teresa Torres continuous discovery)

Prompt 25 - Quarterly business review (QBR) builder

Section 6 - Metrics & North Star prompts (prompts 26–30)

Prompt 26 - North Star Metric selector

Prompt 27 - Input metric tree

Prompt 28 - Cohort and retention analyzer

Prompt 29 - Quality and counter-metric guardrails

Prompt 30 - Executive dashboard wireframe

Section 7 - Communication & alignment prompts (prompts 31–35)

Prompt 31 - All-hands narrative

Prompt 32 - Sales enablement one-pager

Prompt 33 - Recruiting pitch and JD

Prompt 34 - Investor one-liner and 5-sentence narrative

Prompt 35 - Decision-rule generator

Comparison table: 35 prompts at a glance

People Also Ask: 10 questions founders actually ask

1. What is a product vision document?

2. How long should a product vision doc be?

3. What’s the difference between a vision and a mission?

4. Should the founder write the vision alone, or with the team?

5. How often should the vision be updated?

6. How is this different from a product strategy doc?

7. What is a North Star Metric, and do I need one?

8. Can ChatGPT really write a good product vision document?

9. How do I get my team to actually use the vision doc?

10. What are the most common product vision mistakes?

14-day “vision reset” workshop agenda

Common mistakes to avoid (the ones I’ve seen kill vision docs)

Final word: your vision is a bet, not a slogan

SuperFresh AI

43 ChatGPT prompts for non-native English speakers to polish interview answers

41 ChatGPT prompts for SaaS founders in San Francisco to map local partnership opportunities

How to Detect AI-Generated Content

What Is the Best AI Tool for Writing?

AI Newsletter Writing Guide

Get practical AI insights in your inbox