How to measure brand sentiment in AI answers

Brand sentiment in AI answers is how ChatGPT, Gemini, Perplexity and Google's AI describe a business when they name it. How to measure it and correct errors.

Brand sentiment in AI answers is how an engine describes a business when it names it: favourably, neutrally or critically, and with which facts. Measure it by collecting the answers that name the business, reading the tone of each, and listing the claims they make. Then check those claims, because an answer can be warm and still be wrong.

Sentiment only exists where you are named

An answer that leaves a business out says nothing about it. So sentiment is counted over the answers that name the business, and visibility is counted over all of them. Read the two together: a business named rarely and warmly has a different job to do from one named often and described as expensive.

What to record for each answer

  • The tone: favourable, neutral or critical.
  • The claims: what the answer says the business does, charges and is good or bad at.
  • The comparison: who it is set against, and who comes out ahead.
  • The sources: which pages the claims came from.

Wrong facts matter more than tone

AI answers repeat what they find. An old price, a closed branch or a service you no longer offer will be stated with confidence if a page still says so. List every factual claim and mark the ones that are wrong. To correct one, fix it at the source: state the right fact plainly on your own site, on the page the answer cites, and make it agree with every other page.

A score needs enough answers behind it

One critical answer is not a trend, and answers change: in Inciteful's sample a business named in one answer was missing from the next in 11.8% of cases. The figures are in how often AI answers change. Inciteful shows a sentiment score from 0 to 100 only once at least five answers carry evidence of tone, and shows it as unmeasured before that.

The engines describe a business differently

Each engine draws on its own sources. Across the ten pairs of engines in Inciteful's sample, the sites both cited averaged 8.3% of the sites either one cited. A fair description on ChatGPT does not tell you what Gemini or Google's AI answers say, so read sentiment per engine.

How Inciteful reports it

Sentiment shows the score for each client, the answers behind it and the facts the engines state about the brand. Mark a fact as incorrect and it goes into the AEO Boost, the fix list for your own AI coding tool, so the correction lands on the client's own pages.

Questions people ask about AI brand sentiment

What is brand sentiment in AI search?

It is the tone and content of what AI engines say about a business when they name it. It covers whether the description is favourable and whether the facts stated are right.

Can ChatGPT say something wrong about my business?

Yes. It repeats what its sources say, and a source can be out of date or mistaken. Read the answer's cited pages to find where a wrong claim came from.

How do I fix a wrong claim in an AI answer?

Correct the source. Put the right fact plainly on your own site, on the page the answers cite, and keep it the same everywhere else on the site. Then check the same question again over several answers.

Is sentiment the same as visibility?

No. Visibility is how often you are named. Sentiment is what is said when you are. A business needs both counts to know where it stands.

How many answers do I need to measure sentiment?

More than one or two. Inciteful waits for five answers with evidence of tone before it shows a score, and reports the figure as unmeasured until then.

How do AI models describe my brand?

Ask them your buyers' questions and read the answers that name you. The free dashboard does this for ten questions on five engines for any site.

All answers

Your clients’ buyers are asking AI right now. See what it tells them.

A free AI visibility dashboard for any client’s site, on five engines. No card.