Accuracy score
An accuracy score is the share of what AI answers say about your brand that your own facts confirm.
An accuracy score tells you how often AI platforms get you right. It looks at the claims answers make about your brand, such as your prices, where you deliver or which plans you offer, and works out what share of them are true.
Wrong claims cost you quietly. If an answer tells someone you don't deliver to their city, they move on, and you never find out why.
How it's worked out
You can only call a claim right or wrong if you have something to check it against. So an accuracy score starts from your facts: the things that must be true about your business. Each claim then gets one of three verdicts: supported by a fact, contradicted by a fact, or not covered by any of your facts.
The score is the supported claims divided by the supported and contradicted claims together. Claims your facts don't cover are left out, because there's nothing to check them against.
For example, say this week's answers make 50 claims about you. 24 are supported, 6 are contradicted, and 20 aren't covered by your facts. Your accuracy is 24 out of 30, or 80%. The 20 uncovered claims don't count either way.
Accuracy in AnswerOrbit
In Claims you add up to ten facts, and we check every claim an answer makes about you against them. We only mark a claim as contradicted when we can quote the sentence that says it, and you always see that quote.
Accuracy shows how often the answers get you right. Inaccuracy drivers shows the pages behind a wrong claim, so you know which page to get fixed.