Description: What is it?
Cookin’ is a company-wide decision platform. Each company gets a private database of its goals, data and past decisions. Cookin’ flags when stakeholders’ requests collide and scores the competing proposals against company goals. Project tools track work after priorities are set; Cookin’ is where they get set.
Problem: What problem is this solving?
Cross-functional leads get conflicting requests from Sales, Eng and leadership, scattered across tools. With no neutral way to weigh them, the loudest voice wins or the debate stalls for weeks.
Why: How do we know this is a real problem?
- Our interviews: 5 of 17 PM interviewees raised it unprompted. They patch it together with Slack, Asana and Notion, and self-built AI agents, none of which set priorities.
- It’s widespread: 39% of product teams name it as their top challenge (Diginomica).
- It’s expensive: 61% of managers say half their decision time is wasted, about $250M a year at a typical Fortune 500 company (McKinsey).
Success: How do we know if we’ve solved it?
Targets for an 8-week pilot:
- 70%+ of flagged conflicts confirmed as real,
- Time to decision cut in half,
- Under 10% of decisions reopened within 30 days, and
- 3 of 10 pilot companies upgrade to a paid tier.
Kill threshold: fewer than half of the flags are real.
Audience: Who are we building for?
User: cross-functional team leads. Buyer: Head of Product/ COO. Stakeholders join free.
- SAM ≈ $580M: 11.1M US managers x 35% cross-functional leads x 25% at companies on Slack + Asana x $600 a year
- SOM ≈ $6-12M ARR: 1-2% of SAM, about 10-19k seats
What: Roughly, what does this look like?
- Company database: private and isolated, fed by uploads, Slack and Asana.
- Conflict radar: flags requests that collide.
- Taste test: scores competing proposals, with simulated outcomes and confidence ranges.
- Decision record: sends the decision to Jira/ Linear and checks it against results at 90 days.
Business model: Who pays, how, and why?
Companies pay monthly per team lead, and stakeholders join free. Four tiers, each with a monthly analysis limit: Free (5), Team ($50/seat, 50), Business ($100/seat, 200, adds simulations and SSO), and custom Enterprise. Extra analyses cost $10 per 25. One stalled decision costs about $1,350, more than two Team seats for a year.
How: What is the experiment plan?
- Can AI spot real conflicts? Wizard of Oz test on 5 interviewees’ Slack and Asana requests. Fails if under 70% of flags are real.
- Does side-by-side scoring settle conflicts? Prototype test on v0 with live team debates. Fails if under 3 of 5 are decided within a week.
- Will companies share data and pay? Concierge pilot on Free, then an offer to upgrade to Team. Fails if under 2 of 3 share real data.
When: When does it ship?
Tests run Nov 10-23, 2026, and the radar must hit 70% to stay in v1. We build v1 Nov 24-Dec 18, pilot with 10 companies Jan-Feb 2027, and convert pilots to contracts in March, when 3 of 10 must sign.
Readings
- Kevin Yien’s “Non-Goals” in What: v1 ships 4 of its 11 planned integrations and leaves out automatic decisions and scoring people.
- Steve Morin’s focus on success criteria and risks in Success and How: set targets and a kill threshold before building, and tests name the result that would prove us wrong.
