SolutionsFor product teams
Roadmap betswith receipts attached.
Give every proposal the same structure, the same scorecard, and the same standard of evidence. The debate moves from whose deck is more persuasive to which claim is actually supported, and the winner goes on to screens and a working prototype in the same project.
- Owner, admin, member, and viewer roles
- One scorecard across every proposal
- Decisions recorded with rationale
The situation
Planning season rewardsthe most confident slide.
Product teams weigh dozens of proposals with uneven research behind them. The loudest case often wins, and six months later nobody can reconstruct why.
Apples and oranges
Each proposal arrives in its own format and framing, so comparing them means rereading everything and trusting whoever summarized it.
Evidence that goes missing
The research behind a decision lives in slides, docs, and threads that rarely survive a reorg or a change of owner.
Prototypes that forget the research
By the time a bet reaches design and engineering, the evidence and the non-goals live in someone else’s doc, and the prototype drifts from what was agreed.
Hindsight rewrites history
When a bet misses, it is hard to tell whether the call was wrong or the information was, so the team learns less than it could.
The workflow
One pipelinefrom proposal to prototype.
The same steps for every idea, whether it came from a customer call, an executive, or a hack week. The winners continue into screens and a working prototype.
- Step 01
Adopt one scorecard
Start from the default Fast Product Bet scorecard, with nine weighted criteria and six gates. Every concept in the project is judged by the same published rules.
Evaluation - Step 02
Pool the evidence
Research runs and teammates add sources to one library. Claims keep their locators, and contradicting claims are linked so disagreements stay visible.
Evidence - Step 03
Compare structured concepts
Proposals become genomes with the same fields: user, trigger, mechanism, business model, feasibility, assumptions, and risks. Variants keep their lineage.
Concept genomes - Step 04
Decide as a portfolio
Shortlist, validate, build, hold, or reject, with the reason recorded beside the AI recommendation. Experiment outcomes feed a fresh evaluation instead of editing the old one.
Portfolio - Step 05
Design the screens
Mockup Studio plans the flows from the concept’s research and designs each screen per device. Share the prototype link in user sessions before anyone writes code.
Mockup Studio - Step 06
Prototype and review
The Builder builds against the confirmed screens and the handoff, on Next.js, Vite, Astro, or a full-stack monorepo. Engineers review the result as a pull request.
Builder
What matters most
The parts product teamsuse every week.
- PortfolioShortlist, validate, build, or hold.One ranked view of every concept, with status, score, confidence, and gates side by side. Overrides are allowed and recorded, never hidden.
- EvidenceSources and atomic, cited claims.One library for the whole team, with linked contradictions. New joiners can read why a claim mattered without asking the person who found it.
- GovernanceBudgets, provider controls, audit.Roles decide who can run models, record decisions, and manage prompts. Budgets keep exploration inside the quarter’s spend.
- Mockup StudioFlows and screens, designed in your brand.Screens are designed in a brand kit from Brand Studio, so the studio suits new products and concept tests better than additions to an interface you already have.
- BuilderAn AI agent builds the real app, live.A working prototype in a live sandbox, built by an agent against the confirmed screens. Every request shows its cost, and workspace budgets cover all model spend.
- ShipA pull request, a ZIP, or a live preview.Open a pull request on the project’s connected repository so engineers review the code the usual way, or share a deploy preview with its own database for feedback.
Choosing next quarter’s bet from eight proposals, then prototyping it
A product team at a consumer app has eight feature proposals competing for one quarter, and one question from leadership: which of these can we defend with evidence?
An illustrative scenario, not a customer story. Names and numbers are invented to show how the workflow behaves.
Week 1
Each proposal owner turns their pitch into a concept. The evidence library absorbs existing research docs and the results of a fresh research run.
Week 2
Dual evaluation scores all eight. Two proposals show high evaluator disagreement on distribution, which becomes the agenda for the review.
Review meeting
The team moves two forward, holds two pending targeted research, and rejects the rest. One override is recorded with its rationale.
Weeks 3–4
Mockup Studio designs the shared-playlist flows for mobile and desktop from the concept’s research. The prototype link runs in five user sessions, and two screens are refined.
Weeks 5–6
The Builder builds the flow against the confirmed screens. Leadership tries it on a deploy preview, and the code opens as a pull request for the engineering review.
Portfolio
Illustrative data- 78 out of 100Shared playlistsStrong bandShortlist
- 74 out of 100Offline listeningStrong bandValidate
- 66 out of 100Creator tippingHold bandHold
- 61 out of 100Smart notificationsHold bandHold
- 49 out of 100AR album artWeak bandReject
The result
The review spends its time on the two points the evaluators disagreed about, the prototype is built from the screens users saw, and the decision record explains the choice to anyone who joins later.
In the project at the end
- Eight scored concepts
- Decision record
- Screens and specs
- Prototype app
- Deploy preview
- Pull request
Which plan fits
Start with Team.Change plans as you grow.
Team adds shared workspaces, the four roles, project sharing and transfers, deploy previews, and the admin console for providers, prompts, jobs, usage, and audit. It is free during early access; the indicative price after that is $49 per seat per month, billed annually.
Solo
Free
during early access
One workspace, unlimited projects, idea to app.
- 400 image credits a month
- 60 build minutes a month
- No deploy previews
TeamBest fit
Free
during early access
Shared workspaces, roles, and the admin console.
- 4,000 image credits a month
- 600 build minutes a month
- Deploy previews
Enterprise
Custom
annual agreement
Dedicated deployment, managed keys, and procurement support.
- 20,000 image credits a month
- 3,000 build minutes a month
- Deploy previews
Viewers can read the portfolio and its evidence, which suits stakeholders who review decisions without running models.
Take your next planning cyclefrom opinion to evidence.
Invite the team, agree on a scorecard, and bring this quarter’s proposals into one comparable portfolio.
- Solo and Team are free during early access
- Explore a sample project first
- Bring your own OpenAI and Anthropic keys