Skip to main content

How FounderOS reads your idea, and what the number means

Live research, checked twice, scored in the open. Here is every step, and where it stops.

Start free See an example
Placeholder

Demo film coming: a two-minute walk from five questions to a Founder Verdict.

Sources go live as research runs: each of the 73 on the research list is counted the first time a real run returns pages from it.

Seven steps from question to verdict

  1. 01

    You describe it

    Five questions: what you’re building, who for, how it makes money, where you’re starting, and what could kill it. Skip what you don’t know yet; it counts as a blind spot, not a failure.

  2. 02

    Agents fan out

    Up to 10 research agents each take one job — market size, competitors, what buyers say, demand, money, news, rules, sourcing, moat and signals — and search the live web for it at the same time.

  3. 03

    They check each other

    A cross-checker reads every finding. A claim with no link is dropped, and where two agents disagree it says so instead of picking the nicer number.

  4. 04

    It scores the idea

    Six parts are scored out of 100 from the evidence, then weighted and added. Every part comes with one sentence on why.

  5. 05

    The scale

    0 to 100, with 70 as the go line. 0–39 Risky, 40–59 Needs work, 60–79 Promising, 80–100 Strong.

  6. 06

    Six parts, weights shown

    Demand 25% · Market 20% · Competition 15% · Model 15% · Risk 15% · Founder fit 10%. The weights are on every Founder Verdict, so you can redo the sum.

  7. 07

    What it can’t do

    It can’t see private data, talk to your customers, or know your team. It tells you what the public evidence says and what it had to guess, so you know what to check yourself.

What a low score usually means

A low score isn't “give up”. It names the problem, so you can fix it or pick a different idea with your eyes open.

  • Crowded

    Plenty of funded rivals already sell this, and nothing yet sets you apart.

  • Timing

    Interest is fading, or the market isn’t ready to pay yet.

  • Positioning

    The need is real, but the price, buyer or channel doesn’t fit it.

  • Too broad

    “Everyone” isn’t a customer. A narrower first buyer usually scores higher.

An open method you can check

  1. 1

    Claim

    A sentence in your Founder Verdict, like a market size.

  2. 2

    Link

    Its tag: from a source, calculated, or assumption.

  3. 3

    Source

    For a sourced figure, the exact page the search returned, dated.

  4. 4

    Check

    Open it. If it doesn’t say what we said, tell us and we’ll fix it.

Questions about the method

Is the research really live?

Yes. Each run searches the web for your idea at the time you buy it. The sources drawer on every Founder Verdict lists the pages it used and when they were fetched.

Can the score be wrong?

Yes, and it tells you where. The parts with thin evidence are flagged, and the biggest guesses are listed with how to check them. Treat it as a strong first opinion, not a forecast.

Why not just ask a chatbot?

A chatbot answers from memory and rarely shows its work. FounderOS searches first, drops anything it can’t link, and tags every figure so you can see what’s evidence and what’s a guess.

Does it work outside the US?

Yes. Your location changes the market sizing, the currency, the licences checked and the suppliers looked at.

How long does it take?

The Quick Read takes about a minute. A Founder Verdict usually takes 2–4 minutes; you can close the page and find it on your dashboard.

See what the evidence says about yours

Five questions, about a minute, free.

Start free See an example