Mentions and perception

A visibility score tells you how often AI engines name your brand. It does not tell you how they talk about it. Two tabs fill that gap: Mentions & Citations lists the latest answers that named you, and Perception groups everything the answers say about your brand into themes with a tone, so you can see why a score moved and which claims to answer.

#Where to find it

Both are tabs of AEO Insights in the sidebar:

The filter strip above the tabs sets the window (Last 7 / 30 / 90 days) and the market. On Mentions & Citations the brand, engine, topic, persona and Unbranded filters also apply; Perception uses only the window and market — the other controls are greyed out with "Not applied on this tab".

#Mentions & Citations

The card Recent {brand} mentions shows the latest prompts whose answers named your brand. Each row shows the prompt, the logos of the engines that named you, how long ago, and your average position (for example #2).

Click a row to open the answer drawer. It shows:

• Position, Mentions and Runs for that prompt;
• the confidence interval and sample size, for example "Confidence interval 31%–58% · 24 runs";
• an excerpt of the answer where your brand appears.

New mentions appear as each day's measurement completes. If nothing named you in the window, the card says "No answer named your brand in this window."

#Perception

#Tone cards

Four cards open the tab:

Below the cards, one line states the main finding: the most repeated negative claim and how many times it appears, or — if there is no negative claim — your strongest point in the answers.

#Tone trend and theme map

• Tone trend (weekly) plots the positive, neutral and negative shares across the window.
• Sentiment theme map draws each theme as a tile: the area is how often the answers say it, the colour is whether it helps (green) or hurts (red) your brand. A big red tile is the finding. Click a tile to open its answers.

#Theme table and theme detail

The Theme table lists every theme with its Mentions count and Tone (−1…+1): −1 is wholly negative, +1 wholly positive. Click a theme to open its detail below the table: "“{theme}” — N answers", answer by answer — which prompt, which engine, which day and which sentence — plus Prompts that trigger this theme.

Themes are grouped with your market's language rules, so "expensive" and "very expensive" land in the same theme rather than two.

#How your brand is mentioned

This section shows the six most frequent themes as cards, each in the engines' own words with the engine and date of every quote.

On a negative theme you will see Suggested work: publish a page that answers this claim. with an Open task button. It creates a content task under Automation › Tasks. See Content.

#Flag a claim as wrong

If a theme states something false about your brand:

1. On its card, click Incorrect information.
2. Optionally type What is correct.
3. Click Save. The card then shows flagged.

Flagged claims are listed in Incorrect information flags with their status. Click Work it on the Truth screen to move the claim through review and correction. See Misinformation.

#Sources appearing alongside negativity

This table lists domains that are frequently cited in negative answers about you, with the Negative share and the total Answers count. A domain needs at least 3 answers behind it to be listed. It shows that a site appears together with negativity, not that it causes it — read the rate and the count together.

#Said about rivals, not about you

This table compares only answers that name both you and a rival: same question, same engine, same day. A difference here is a real difference, not a side effect of a different question mix.

The header shows how many shared answers were compared. Each row is a theme with Said about them (the share of shared answers that said it about a rival), Not said about you (the share that said it about you — both over the same answers), the rivals it was said about and Questions it came from. Go to a brief opens the Content page, where you can brief a page that earns the same words.

If there are fewer than 4 shared answers the table says "Too few shared answers to compare" instead of showing shares.

#Limits and availability

Mentions and perception are included in every plan. They are built from the answers your daily measurement already collects, so they cost nothing extra; coverage grows with the number of prompts and engines you measure — see Plans and limits.

#Common questions

The Perception tab is empty. Themes are extracted from measured answers. A new workspace shows "No perception data yet" until its first answers have been read.

A theme's tone looks wrong. Flag it with Incorrect information or open its detail to read the exact sentences behind it. The tone of a theme is the average across those sentences.

Why does the week on the tone cards differ from the window? The tone cards read the most recent week; the trend chart and the theme table use the whole window you chose.

#Related pages

• Visibility score
• Competitors
• Sources
• Misinformation
• Content