The Flare score answers one question: when AI engines are asked the questions that matter in your category, how well do they present you? Not only whether you appear, but how prominently, how accurately, how evenly across engines, and how favourably. It is a single number from 0 to 100, shown at the top of every brand page, on your dashboard, and on every profile in Explore.
Why one number instead of five
How often you're named, on its own, rewards being talked about. It says nothing about whether you were named first or fifth, whether what was said was true, whether you show up on all five engines or only one, or whether the framing helped you. Two brands named equally often can be in completely different positions, and a leaderboard built on one metric hides that.
Flare combines five measurements brandflare takes from the same set of answers, and always shows them beside the number.
The five factors
Select See breakdown on the Flare card to open Your scores: the five factors behind the number, each scored 0–100, each with a question mark explaining what it measures.
| Factor | The question it answers |
|---|---|
| Visibility | How often do engines name you at all? |
| Prominence | When they name you, how early, how strongly, and how much of the conversation do you take from your competitors? |
| Accuracy | How much of what engines say about you is true? |
| Consistency | Do you appear evenly across all five engines, or lean on one? |
| Sentiment | How favourably are you described? |
Visibility is your mention rate: the share of answers that name you. It is the same figure shown as Visibility on your overview, documented in how visibility is measured.
Prominence separates being the first brand recommended from being the seventh name in a list. It draws on where you land in each answer, how firmly the answer recommends you, and your share of voice against the competitors you track.
Accuracy comes from Answer Accuracy: false claims that are still open count against you, serious ones more than minor ones. Claims you've marked as not a mistake don't count.
Consistency measures evenness, not level. A brand named constantly by ChatGPT and almost never elsewhere scores poorly here, however good its Visibility looks, because most of its customers are being told nothing.
Sentiment reads how the answers describe you, from warm recommendation to caveated mention.
The arithmetic that joins the five into one number isn't published. The factors are the part you can act on, and they're always shown.
What counts, and what doesn't
- Questions that name your brand are left out. An engine naming you when the question asked about you measures nothing. Only questions that don't mention you feed the score.
- Every engine counts. ChatGPT, Gemini, Google Search, Perplexity and Claude all contribute, and the breakdown by engine is always on your overview.
- Failed answers are excluded, not counted as a miss. If an engine was down during a check, that gap doesn't cost you.
The bands
| Score | Band |
|---|---|
| 80 and above | Very high |
| 65–79 | High |
| 50–64 | Moderate |
| 30–49 | Low |
| Below 30 | Very low |
The colour of the score follows the band, and the same scale colours the bars on your overview, so green and amber mean the same thing everywhere.
A perfect 100 is out of reach in practice, and deliberately so: it would mean being named in every answer, first, on every engine equally, with nothing false said and uniformly glowing framing.
Early scores
A brand's first check gives it a score, but one check is a small sample. Until enough answers are in, the score is labelled Early score, still settling, and it is treated with caution, so one lucky or unlucky check can't produce a 97 or a 12. The label goes once the evidence is there.
When a factor wasn't measured
A factor with nothing behind it doesn't score zero, and it doesn't score a polite 50. It is left out, and the score is built from the others. The breakdown marks it as not measured.
The usual case is Accuracy on a profile in Explore. Brands in the public index are checked for mentions but not against a fact sheet, so nobody has looked for false claims, and "no false claims found" would be an unearned compliment. Track the brand yourself and Accuracy is measured from the first check. This is also one reason a brand's score in your workspace can differ from its public profile: your questions, your competitors, and a full accuracy check.
It's absolute, not a ranking
Your Flare score moves only when your own results move. A competitor improving doesn't lower it; a weak category doesn't flatter it. That is what makes the trend line mean something, and why the score works as a target rather than a league position. Rankings exist too, on Explore and in Competitors, but they're built from the score, not the other way round.
Reading your score
There is no universal "good" score: a household name and a two-year-old challenger live in different ranges. Three readings matter more than the number itself.
- Direction. The Flare card shows the last 90 days. Is the line rising after the work you shipped?
- Which factor is dragging. A brand with high Visibility and low Accuracy has a completely different problem from one with the reverse. The first needs its sources corrected; the second needs to be talked about more.
- The gap. How far you sit from each competitor you track, on the same questions and the same answers.
Every number has receipts
Every answer behind the score is stored. From the score you can open the factors, from a factor the questions and engines behind it, and from there the answer itself, word for word. If a number surprises you, the reason is a few clicks away, and Recommendations turns the weakest factors into a list of what to do next.