Visibility tells you whether AI mentions your brand. It does not tell you how. Two brands can both be mentioned in the same answer and come away with completely different impressions: one framed as the trusted default, the other as the cheap alternative. That framing, the words an assistant reaches for and the tone behind them, is what Brand Themes measures.
The themes AI uses for you
A theme is a recurring frame the assistants use when they talk about your brand. Across hundreds of answers, certain ideas keep coming up, and SurfacedBy surfaces them as named themes with how much of your coverage each one accounts for. It is the difference between knowing you are visible and knowing what you are visible for.
Each card carries a short description of what the theme means, the share of answers it appears in, and a link to the exact prompts behind it, so a theme is never a black box. You can read straight down the list and see the story AI is telling about you, ranked by how often it comes up.
Sentiment, theme by theme
A single overall sentiment score hides the thing you need to act on. You can be described in glowing terms on one theme and shakily on another, and the average flattens both into a number that tells you nothing. So sentiment is attached to each theme: the split of positive, neutral, and negative answers within that frame.
That is where the useful signal lives. A theme that is central to your coverage but carries a slice of negative answers is a clear, specific problem: a story AI tells often, and sometimes tells badly. A theme that is small but entirely positive might be worth amplifying. One honest framing point: this is AI answer sentiment, how the assistants describe you, not a measure of what your customers feel. It tells you how the machines are characterizing your brand, which is its own thing worth managing.
From a perception problem to a fix
Naming a perception problem is only useful if you can do something about it. A weak or inaccurate theme does not just sit on a chart; it routes into an evidence-backed opportunity, classified by what is actually wrong. When the issue is that AI states your facts incorrectly, that is an accuracy problem with a different fix than a positioning problem where AI simply frames a competitor as the better choice. The diagnosis points at the work.
To keep this honest as your brand changes, SurfacedBy refreshes what it knows about you between full checks and surfaces newly discovered offerings for you to confirm, so the themes reflect the brand you are now, not the one you were six months ago. If AI is getting your story wrong, the brand understanding audit is a good place to see why, and the diagnostic behind why AI recommends competitors covers the positioning side.
A new brand will show a thin set of themes, because the assistants have not read enough to form a consistent story yet. That is labeled as early rather than dressed up as a confident read. As coverage grows, the themes sharpen.
You will find Brand Themes on the AI Visibility page for any tracked brand, with a compact version on the Overview.
A note on timing
SurfacedBy changes often. We build from a mix of customer requests, our own research, and keeping pace as the assistants change how they describe brands, so the themes and sentiment here reflect how the feature looked at launch. The diagnosis that turns a perception problem into an action has gained depth since, and it will keep gaining it.
