The founder’s judgment, in the room — even when they can’t be.
Past a dozen people, a founder can’t sit in every debate. Teams either wait for them (slow) or guess what they’d say (drift). Founder OS is an AI stand-in trained on how a specific founder actually decides — so the waiting becomes unnecessary, and the guessing becomes honest.
How a company uses it
1
Teach it how you decide.
The founder answers seven questions — vision, customer, contrarian beliefs, the lines they’d never cross, how they actually talk. An AI distills this into an operational profile; the founder edits it until it sounds like them.
2
Ask what you’d ask them.
Anyone on the team drops in a stuck decision, a Slack thread going in circles, a half-formed doubt — anything they’d bring to the founder if they could grab ten minutes.
3
Get the take — or the handoff.
The stand-in answers in the founder’s voice, anchored to a stated principle, with its confidence in plain language. When the evidence is thin it doesn’t bluff — it drafts the decision as a one-line question for the real founder.
Try it on founders you know
Pick a company. Each profile below was assembled the way a real workspace is — from how the founder publicly thinks, decides, and talks. The first question plays to what’s documented; the second sits at the edge of it, where the stand-in’s real job is to hand the decision back instead of bluffing.
Brian Chesky — Brian Chesky runs Airbnb the way a designer runs a review — one roadmap, fewer and bigger bets, and the conviction that the company itself is the product. Distilled from his founder-mode talk, interviews, and shareholder letters.
We want to spin up a growth pod with its own roadmap and A/B test our way to a better listing page — fourteen variants ready to go. Half the team says we should just decide what good looks like and ship it.
asked by Priya, Product
✳Airbnb founder · stand-in
Kill the separate roadmap before it kills the product. A thousand A/B tests don't add up to a point of view — they add up to a Frankenstein page nobody designed. Decide what great looks like, ship it, and use the data to verify the call, not to make it for you.
Anchored to
One company, one roadmap, fewer things done excellently — and the belief that the company itself is the product, not a holding company of pods.
The question back
If we need fourteen variants to choose, do we actually have a conviction about what this page is for?
What I’d do
Pick the strongest design in one review this week, ship it to everyone, and measure against the call we made.
I'm sure about this one — it's written into how we decide. · confidence 9/10
The take is anchored to positions the founder has stated publicly — quoted, not invented — with confidence in plain language.
An unofficial demonstration. Public-figure profiles are distilled from interviews, essays, and talks; the answers are illustrative of how Founder OS works and are not affiliated with or endorsed by these companies or founders. Sundial is fictional — and it’s the live workspace behind the gate.
Where it’s heading
Every answer the real founder reviews becomes precedent. Corrections become principles. The stand-in gets sharper the longer the room stays open — the moat isn’t the persona, it’s the accumulated judgment.