Build what wins deals, in the buyer's own words

The gaps buyers name most, ranked by the deals they touch, with the quotes behind them. Roadmap arguments settled with evidence, not the loudest opinion.

Start free trial

Works inand any MCP client

Question: Rank the product gaps that cost us deals in the last two quarters, with the buyer quotes.

Used Calven MCP

SCIM provisioning on the mid tier: 6 lost deals. "Security wouldn't sign off without SCIM." (VP IT, lost to Northwind)
dbt Cloud integration: 4 lost deals. "Half our models live in dbt Cloud. That was the end of it." (Data lead, no decision)
On-premises hosting: 3 lost deals, all to Northwind, 2 of them banks.

Northwind already offers two of the three.

Sources: 21 win/loss interviews · 38 calls · Northwind dossier

Example workspace. Names and numbers are illustrative.

What can product teams use Calven for?

Roadmap cases, PRDs, parity calls and launch messaging that start from the deals you lost and the words your buyers use on calls. Calven is the approved context your AI workflows need to scale across the team and still be trusted.

Use cases

Example prompts for gap prioritization

Ask any of these in Claude, ChatGPT or any AI tool connected to Calven.

Rank the product gaps that cost us deals in [the last two quarters], with the buyer quotes.
Check how far [gap] reaches beyond lost deals: calls, open pipeline, trend.
Does [competitor] have [gap], and did it decide the deals we lost to them?
Turn this gap shortlist into a one-pager for roadmap planning: [paste the shortlist]
Tell sales what happens with each gap and what to say meanwhile: [paste the decisions]
Backtest our gap ranking on last year's deals and find out which rule actually predicts the losses that follow.
Replay every deal we lost on [gap] as if we'd shipped it, and count how many would really have flipped.
Is [gap] real or one loud customer? Give me a Bayesian estimate and update it as each new deal lands.
How many lost deals named [gap], and how much was at stake?
Did [gap] decide those deals or come up in passing?

Open the full prompts on GitHub

Example prompts for roadmap input

Ask any of these in Claude, ChatGPT or any AI tool connected to Calven.

Build the evidence card for [candidate]: who asked, deals lost, who else has it.
Build evidence cards for [next quarter]'s shortlist: [paste the shortlist]
Does [candidate] fit our strategy, or only the needs of [the account asking]?
Make the strongest evidence-based case against building [candidate].
Write next quarter's roadmap narrative, with why we said no: [paste the items]
Run the [Q3] roadmap retro: did what we shipped move the deals? [paste planned vs shipped]
Build a cost-of-delay spreadsheet for my roadmap shortlist, using real deals at stake, and sequence it by CD3.
Stage a judged debate between [candidate A] and [candidate B] for the last roadmap slot, with a ruling and a dissent.
Place my roadmap candidates on a Wardley map and tell me where to differentiate, match cheaply or buy.
How many deals named [candidate], and how many were lost?

Open the full prompts on GitHub

Example prompts for competitive parity

Ask any of these in Claude, ChatGPT or any AI tool connected to Calven.

Map where [competitor] is ahead of [product] and whether buyers care.
What do buyers say about [capability] in deals against [competitor]? Quote them.
If we match [competitor] on [capability], do we lose what wins us deals today?
Write the parity decision memo for [competitor]: [paste the decisions]
Give sales the answer for when a buyer asks about [capability] from [competitor].
Rebuild the weighted scorecard a buyer uses to pick between us and [competitor], and find the weight where [capability] flips it.
Build a payoff matrix of our parity options against [competitor]'s likely responses and find the move with the safest worst case.
Run a Kano survey on our personas to sort sales' parity requests into must-haves, delighters and noise.
Where is [competitor] ahead of us according to their dossier?
Did any buyer say they chose [competitor] for [capability]?

Open the full prompts on GitHub

Example prompts for launch positioning

Ask any of these in Claude, ChatGPT or any AI tool connected to Calven.

Frame the [feature] launch on one page: pillar, persona, pain, proof.
Write the [feature] launch messaging for [persona] from this frame: [paste the frame]
Fact-check the [feature] launch copy against the product brief: [paste the copy]
Read this announcement as [persona] and tell me what they would doubt: [paste it]
List every published document the [feature] launch makes stale in [area].
Write the sales FAQ for [feature], including what it does not do.
Run a conjoint test of my launch headlines, benefits and proof with our personas to find which element does the work.
Play [competitor]'s product marketer and write the counter-campaign they'd run the week we ship, then patch our copy.
Forecast 90-day adoption for [feature] from our past launches as a reference class, adjusted for real demand evidence.
Which pillar does [feature] belong under in our messaging?

Open the full prompts on GitHub

Example prompts for opportunity framing and PRDs

Ask any of these in Claude, ChatGPT or any AI tool connected to Calven.

Frame [problem] for a PRD: who has it, how often, what it costs us.
How do customers handle [problem] today? Quote the workarounds.
Is there a market trend that makes [problem] urgent this year?
Propose success signals and the riskiest assumptions for [opportunity].
Pressure-test this PRD's problem section against the evidence: [paste it]
Map the push, pull, anxiety and habit forces on [problem] from customer quotes, so the PRD knows what stops people switching.
Size [problem] bottom-up from accounts we can name, as a P10 to P90 range, and show which input matters most.
Pull every assumption out of my PRD, plot importance against evidence, and write the cheapest test for the riskiest three.
Give me five quotes about [problem], with who said them.
Did [problem] decide any deals? Won or lost?

Open the full prompts on GitHub

Example prompts for customer feedback synthesis

Ask any of these in Claude, ChatGPT or any AI tool connected to Calven.

Write the [last quarter] feedback synthesis for [product]: themes, sentiment, quotes.
Break [theme] down by persona, segment and account, with quotes.
What changed in customer feedback between [last quarter] and [this quarter]?
Give me the full picture on [theme]: mentions, deals, momentum, quotes.
Join what each account said to my renewal data and find which themes predict churn, not just which are loudest.
Reweight my feedback synthesis for the accounts we over- and under-hear, and show which themes change rank.
Put every rising theme on a control chart and tell me which are a real shift and which are normal noise.
What are the top five customer pains this quarter for [product]?
Which themes grew most since last quarter?
Give me three quotes under [theme] with who said them.

Open the full prompts on GitHub

Example prompts for competitive teardowns

Ask any of these in Claude, ChatGPT or any AI tool connected to Calven.

Which competitor deserves the next teardown, based on [last quarter]?
Build the full teardown of [competitor] for [last quarter].
Compare our product and pricing with [competitor] side by side.
Check the [competitor] battlecard against this teardown: [paste the teardown]
Model our next move against [competitor] as a payoff matrix and find the option that holds up whatever they do.
Write the battlecard [competitor]'s PMM uses against us, then show me where it beats ours.
Map every segment on a fight-or-walk 2x2 against [competitor] and give sales rules of engagement for qualification.
What did [competitor] do in the last 90 days?
Why do we lose to [competitor]? Quote the buyers.
Where is [competitor] ahead of us on features, and do buyers care?

Open the full prompts on GitHub

Example prompts for pricing input

Ask any of these in Claude, ChatGPT or any AI tool connected to Calven.

How did buyers rate the price of [product] in [last quarter], by competitor?
Compare our pricing with [competitor]'s, including moves in [the last six months].
Which packaging questions keep coming up in deals for [product]?
Write product's input note for the [Q4] pricing review: [paste the evidence]
Have our buyer personas answer a Van Westendorp price ladder for [plan] and find the acceptable price range before the review.
Model whether moving [capability] to a higher tier pays for itself, and find the upgrade rate where it breaks even.
War-game a [competitor] price cut over three rounds and find the response that holds up whatever they do next.
What share of buyers found us pricier than [competitor] this year? With the sample.
Which deals did we lose on price in [window], and to whom?
Which capabilities do buyers expect in a lower tier? Quotes.

Open the full prompts on GitHub

Example prompts for sales FAQs

Ask any of these in Claude, ChatGPT or any AI tool connected to Calven.

Build the sales FAQ for [product] from the approved product brief.
Add the answers reps need when a buyer compares us with [competitor].
Fact-check this FAQ against what changed in [the last six months]: [paste the FAQ]
Write a 40-question eval set, with gold answers from our product brief, that catches AI tools overclaiming to buyers.
Play a tough [persona] in a fifteen-minute role-play, grade each rep reply against the product brief, and score the drill.
Does our product do [capability]? What does the brief say?
What does the brief say we do not do?
Which questions came up as capability objections in lost deals this quarter?
Which question do buyers ask most that our reps answer differently from the product brief?
Which known weakness comes up most in lost deals, and what's the honest answer?

Open the full prompts on GitHub

Example prompts for release notes

Ask any of these in Claude, ChatGPT or any AI tool connected to Calven.

Write the [release] notes for [personas] from this list: [paste what shipped]
Which published documents does [release] make stale in [areas]?
Write the deprecation notice for [capability], removed on [date].
Fact-check these release notes against the product brief: [paste the notes]
Score every account affected by removing [capability] for churn risk, and tell me who needs a call before the notice.
Design an A/B test for the release email, messaging words versus customer words, with a power calculation on my list size.
What problem does [feature] solve, in a customer's words?
Is "[claim]" in the product brief?
Who relied on [capability], according to customer quotes?
Which customer quotes asked for what we just shipped, and from which accounts?

Open the full prompts on GitHub

Example prompts for feature naming

Ask any of these in Claude, ChatGPT or any AI tool connected to Calven.

Propose names for [feature] from the words customers use for what it does.
Test these three names as [persona]: [paste the names]
Does the name [name] still match what customers call it?
Run a pairwise naming tournament judged by our personas and rank my candidate names with a Bradley-Terry model.
Attack my shortlisted names as a competitor's rep would, check collisions with their features, and see which survive.
What do customers call [capability] on calls?
Does any competitor have a feature called [name]?
How would [persona] read the name "[name]"?
What words do buyers in lost deals use for [capability] that buyers in won deals don't?
Which of our feature names never appear in customer quotes?

Open the full prompts on GitHub

Example prompts for beta recruiting

Ask any of these in Claude, ChatGPT or any AI tool connected to Calven.

Build a beta invite list of [12] accounts in [segment] for [feature], with the reason each.
Write the beta invitation for [persona] in their words, for [the beta scope].
Summarise what the [feature] beta accounts said over [six weeks].
Run a power calculation for my [feature] beta and pick a stratified sample of accounts that asked and accounts that didn't.
Build a spreadsheet with live formulas that scores every [segment] account for my beta, with a weight sensitivity tab.
Which customers asked for [capability] on calls or in deals?
Who is the champion at [account]?
Which accounts mentioned [problem] with negative sentiment?
Which accounts asked for [capability] and also have an open renewal deal?
Which segments have no account that raised [problem]?

Open the full prompts on GitHub

Example prompts for advisory board prep

Ask any of these in Claude, ChatGPT or any AI tool connected to Calven.

Draft the [Q4] advisory board agenda and pre-read from the evidence.
Propose advisory board members across [segments], with why each belongs.
Write discussion questions on [topic] that get direction from [personas].
Run a two-round Delphi panel of my board members, built from their accounts' calls, to find where they'll split before the session.
Put odds on the beliefs my roadmap rests on, then update them like a Bayesian with what the board says.
Which strategic themes grew most this quarter?
Which product gaps have the most deals at stake?
What did advisory board members say on calls since the last session?
Which themes do our largest customers raise that the rest of the base doesn't?
Which advisory board accounts have an open renewal or expansion deal?

Open the full prompts on GitHub

Example prompts for discovery interview prep

Ask any of these in Claude, ChatGPT or any AI tool connected to Calven.

Write a discovery interview guide for [persona] on [outcome].
Brief me on [contact] at [account] before the interview.
Rehearse this interview with me as [persona]: [paste the guide]
Code my interview notes into an opportunity tree twice, independently, and measure where the two codings disagree.
Draw a saturation curve per segment from the calls we have and tell me how many more interviews each needs.
s jobs to be done?",
,
Which of [persona]'s pains on the canvas have no customer quote behind them?
What workarounds have customers described for [job]?
Which lost deals mention [problem], and what did the buyer choose instead?

Open the full prompts on GitHub

No new tool to learn

Calven works where your team already does: in Slack, and in Claude, ChatGPT or any AI tool through Calven MCP. Nobody opens another app to get the approved story.

In Slack

Mention @Calven in a thread or send it a DM. It answers from the approved version, with sources.

Nina Volkova9:12 AM

Does Northwind have a dbt Cloud integration, and did it decide the deals we lost to them?

CalvenAgent9:12 AM

Yes, since June, per the Northwind dossier.

It decided one deal. dbt Cloud came up in 4 lost deals: 1 went to Northwind, 3 ended in no decision. Those buyers didn't switch, they stalled.

In their words: "Half our models live in dbt Cloud. That was the end of it." (Data lead, no decision)

Sources: Northwind dossier · Win/loss, 4 deals · Deal drivers

Example workspace. Names and numbers are illustrative.

In your AI tool

Add Calven to the AI workflows you already run. Every step reads from the approved version, with sources.

Question: Pressure-test this PRD's problem section against the evidence.

PRD: SCIM provisioningGoogle Doc

Used Calven MCP

Holds: the cost. SCIM decided 6 lost deals this year, 4 to Northwind, and security sign-off is the reason buyers give.

Push back on: "every enterprise prospect asks for SCIM". It came up in 9 of 38 calls, all mid tier.

Missing: the workaround. Two customers script provisioning by hand today, and Kestrel Payments said so on a call.

Sources: 21 win/loss interviews · 38 calls · Northwind dossier

Example workspace. Names and numbers are illustrative.

What changes when product teams' AI tools read from Calven?

Product teams using AI without a shared source get confident PRDs built on anecdotes. Connected to Calven, every roadmap case, PRD and launch starts from the same deal evidence and approved story.

Pain point	Without Calven	With Calven
Prioritization	Roadmap debates won by the loudest voice or the last big customer	Gaps ranked by the deals they cost, with the buyer quotes behind each
Parity	Feature parity chased on every competitor launch	Matched where buyers decided deals on it, skipped where they did not
PRDs	Problem sections that start from "several customers asked"	Problem sections with quotes, affected accounts and the deals at stake
Feedback	Customer feedback scattered across ten tools and read by whoever has time	Themes with quotes, momentum and revenue tie, every period
Launches	Launch messaging written after the build, claims checked by nobody	Positioned from the approved story on day one, every claim checked against the product brief
Sales questions	Reps ask product the same questions in Slack every week	A sales FAQ answered from the product brief, including what you do not do

Two weeks trial. You'll only need a day.

A day is all it takes to see what the agents can do for you.
You won't go back.

Start free trialNo credit card required

Questions, answered

How does Calven help product managers?

It turns win/loss interviews and calls into evidence product can use: gaps ranked by the deals they touched, competitor feature comparisons, customer themes with quotes, and launch messaging from the approved positioning.

Can Calven rank feature requests by revenue impact?

Calven ranks product gaps by the number of deals each one touched in win/loss interviews and calls, with the buyer's words and the competitor involved. It uses the deals in your workspace, not estimates.

Does it work in Cursor or Claude Code?

Yes. Calven MCP works in any MCP client, including Cursor, Claude Code, Codex and VS Code with GitHub Copilot, so a PM or engineer can pull buyer evidence while writing a PRD or a spec.

Do product teams need to log into Calven?

No. Product marketing runs Calven. Everyone else reaches the same approved context from the AI tool they already use over Calven MCP, or in Slack. Professional and Enterprise include unlimited seats, so each person signs in as themselves.

Which AI tools work with Calven?

Any client that supports remote MCP: Claude, ChatGPT, Cursor, VS Code with GitHub Copilot, Claude Code, Codex and others. Paste https://api.calven.ai/mcp, choose OAuth and sign in.