The bar we will not publish below.
Someday soon, Jinn will show you a number for how AI sees your brand. Before you ever see that number, it has to clear a bar we set for ourselves — and it hasn’t cleared it yet, so you won’t find it here. This page is the method behind that refusal: the bar, the checks, and why we’d rather show you nothing than a number we don’t yet trust.
There is no accuracy figure on this page, and that is deliberate. The discipline below is why.
What this is
The method behind measuring how AI sees a brand: the bar a number must clear, and the separate checks that catch the ways measurement goes wrong. A method, not a score.
The floor
Before any score reaches you, our automated grader’s verdicts must agree with a human-labeled golden score at least 85% of the time. Eighty-five percent is the bar to clear — not a result we’re claiming.
Why no number
The method is real, working code; the grader hasn’t yet cleared its own bar, so nothing publishes — not on this page, and not in the product. That refusal is the point.
The bar we hold ourselves to.
The rule is simple: our grader’s verdicts must agree with a human-labeled golden score at least 85% of the time, claim by claim, before any score is allowed in front of you. And the same code that computes that agreement in our tests runs the live gate — so the bar can’t quietly drift away from the check that enforces it.
Eighty-five percent is the bar, not a result. It’s the line the method has to clear before a number is fit to show you — never a claim about where the grader stands today. We may raise it; we will not quietly publish under it.
Each check catches a different failure.
A single agreement number can be fooled, so the discipline is more than one check. Each one below catches a different way measurement lies.
- agreement
- Our verdicts must match a human-labeled golden score at least 85% of the time before any score publishes.
- recall
- AI engines fail in correlated ways — several engines can agree on the same wrong fact. So a separate check verifies how many known-bad claims got caught. High agreement with low recall is treated as a failure, not a pass.
- coverage
- A brand is only scored when at least five different AI engines answered every question. Thin data doesn’t get a score — the run goes back in the queue.
- confidence
- A published score will carry its honest margin of error. When there isn’t enough data to estimate honestly, the method says “no estimate” instead of guessing.
- no accusation
- Calling an answer “wrong” requires an authoritative source. Without one, the verdict is downgraded to unverified disagreement — the method never accuses on a hunch.
These checks are deliberately separate. A run can pass one and fail another, and a single strong number never buys its way past the rest.
This is the discipline behind the numbers. The place to start is your own brand. Read your brand free
Publishing nothing, on purpose.
You’ve probably noticed that most AI tools lead with a big accuracy percentage. We could put one here too. We don’t, because the grader hasn’t yet cleared the bar it has to clear, and a number below the bar is a number you shouldn’t be shown.
When it’s good enough to trust, you’ll see it. Until then, nothing is the honest answer — and the more defensible one.
Keep exploring how Jinn works.
This page is the measurement discipline. The full tour shows the mechanism it sits inside — how Jinn learns, wears, and works a brand.
- The full tourThe mechanism end to end: how Jinn learns, wears, and works a brand.
- The recordThe 346-signal Brand DNA record, group by group.
- Models & routingWhich model does which job, and why you never have to care.
- Guardrails & spendThe fixed rules that guard your money and your name.
- Check it yourselfWhat “verified” actually means here: human approvals, recorded history, and facts that age on purpose.
Measurement, answered.
- Why doesn’t this page show an accuracy number?
- Because the number is not yet trustworthy. Our grader must agree with a human-labeled golden score at least 85% of the time before any score publishes; until it clears that bar, we publish nothing rather than a figure we can’t stand behind.
- Is 85% your accuracy?
- No. 85% is the bar the method has to clear before a result is fit to show you — not a measurement of where the grader stands today. A bar, not a score.
- How do you know the AI engines are not agreeing on the same wrong answer?
- That exact failure is why agreement isn’t the only check. A separate recall check verifies how many known-bad claims got caught, because engines fail in correlated ways — high agreement with low recall counts as a failure, not a pass.
- Can I measure my brand’s AI visibility in Jinn today?
- What’s live today is Fama: it shows you what the AI engines actually say about your brand, starting with the free AI-visibility audit. The published score this page describes is the part still behind the bar — it ships only once the grader clears its floor.
We would rather publish nothing than a number we do not trust.
That is the whole discipline. Start with the free brand read — the part of Jinn that is ready for you today.