How we form your Brand Score

1. Your Brand Score at a glance

Your Brand Score answers two questions every business cares about when customers ask AI assistants for recommendations:

• 75–100% — Visibility. Do the AI engines name you, name you early, and cite your website? Visibility keeps every percentage point not backed by eligible perception data.
• Up to 25% — Perception. How do they position you against your competitors? Each of five dimensions contributes 5% after at least three successful evaluations.

Both come from the same place: real answers, stored and traceable. Nothing is estimated.

2. We ask real questions to real AI engines

Each measurement round, we take your catalog of questions — built from your products, services, and how customers actually search — and ask them to each AI engine in your plan (ChatGPT, Google AI Overviews, Gemini, Claude, Perplexity), with web search on, exactly as a real user would. One question asked to one engine is one measurement.

Every answer is stored, so any number we show you can be traced back to the literal text that produced it.

3. Visibility: an answer earns partial credit — not a yes/no

For every answer, we grade what actually happened to your brand — because being the first recommendation and being a footnote are not the same thing:

Your Visibility score is the average across all of your measurements — no single question can dominate it.

4. Your score is a standing, not a snapshot

AI answers naturally differ from day to day — ask the same question twice and you can get different recommendations. That is normal, and it is exactly why your score is the weighted average of your recent runs — the most recent ones count more — rather than any single run.

Your first run starts the score. From the second compatible run onward, every available result contributes; the baseline is marked as establishing until three runs have been completed.

You wouldn't change your SEO because of one search result. One run is a photo; several runs are a video. Ordinary back-and-forth cancels out, while a genuine change — new competitors, better content — shows up as a steady movement within a few runs.

5. Perception: how the engines talk about you

Alongside your questions, every round asks a set of unbranded ones — questions like "Which firms are most trusted for this?" — so the engines reveal, in their own words, where you stand against your competitors. Each answer is scored across five stages:

Each stage is graded with partial credit — the same principle as Visibility: being called the clear leader counts for more than being listed among many, and winning a head-to-head outright counts for more than a tie. Your Perception score is the average across the stages. Each sufficiently measured stage contributes 5% of your Brand Score, up to 25% when all five qualify; one stage never occupies the whole perception allowance.

6. What the numbers mean

100 means your brand is named first and cited in every single measurement — perfection. In practice, scores live in bands: presence (named somewhere), position (named early), authority (named first and cited). We show you where you sit and how far the next band is, rather than implying a single point of precision.

7. When your setup changes, the baseline restarts

If you edit your catalog — add or remove questions, pin new competitors — or your engines change, your recent history measures something different from today. The averaging window starts over, and the app tells you so.

Comparing "before you added 10 questions" to "after" isn't a trend — it's two different surveys. A reset isn't bad news; it's the score telling the truth about what it can and cannot compare.

8. Two things we will never do

• We never count "couldn't measure" as "you weren't visible." If an engine couldn't complete an answer, we exclude it and show how many measurements your score is built on.
• We never show you a number we can't explain. Every score comes with what it is built on, and every claim on this page maps to a formula in our product.