Help centerMonitor
Answer audit
Your home page: brand visibility, the sources AI uses, and every answer with the brands it names.
Answer audit is the first page you see. Its top half summarises how AI treats your brand over the selected range. Below that, All chats lists every answer, with the most important ones first.
The summary line
Under the title, one sentence gives the headline numbers for the range:
- AI recommends you in X% of answers that suggest tools. Needs at least 5 answers that recommend a product. Until then the line shows You appear in X% of answers, which needs 10 answers.
- Named first in X% of mentions. Needs at least 3 mentions.
- Top competitor. The most visible competitor with at least 10 answers. Click its name to open Insights.
A part is left out until it has enough answers to be meaningful.
Visibility
This card charts the share of answers that name each brand, day by day. Your brand is the thicker blue line. It plots the same brands as the Brand visibility table: the top seven, plus yours.
- D W M group the points by day, week or month. An interval finer than the range allows is greyed out. A week or month only partly inside the range says which days its point covers.
- The two icons switch between Trend view and Comparison view. Comparison view is a bar per brand for the latest period.
- Hover the chart to see every brand’s value on that date.
- An amber dashed line with a diamond marks a day when the AI models changed. Hover it to see which models were added or removed.
- A hatched Not read yet band means the answers for those days exist but haven’t been read yet. See Scans and data freshness.
Brand visibility
This table ranks brands for the selected range and filters. The numbers are the same as on Insights. Your row is tinted blue. The expand icon opens every tracked brand full screen.
- #
- Rank by visibility. Ties share a rank. A small arrow shows whether the brand moved up or down since the previous period; hover it for the previous rank.
- Visibility
- Share of answers that name the brand.
- Recommended
- Of the answers that recommend any product, the share that recommend this brand.
- Sentiment
- How positively answers talk about the brand, from 0 to 100. 50 is neutral. Hidden on phones.
- Position
- Average place in the answer; #1 means named first. Lower is better. Hidden on phones.
Markers next to a number
- ↑4.2Better than the previous period (points)
- ↓2.6Worse than the previous period
- NewNot seen in the previous period
- ↑Moved up in rank (hover for the old rank)
Source signals, Domains and Domain types
These three cards preview the Sources pages for the same range:
- Source signals
- The 6 pages (URLs) AI uses most. In answers is the share of answers that use the page, with New or a change marker. Cited is the share of answers that cite it. Click a title for the page’s details, or All URLs for the full list.
- Domains
- The 8 websites AI uses most, with In answers, Citation rate (how often a use is a citation) and Type. Click a type to see every domain of that type.
- Domain types
- How source uses split by type of site, such as your website, competitors, community, reference or editorial. Click a bar to filter Domains by that type.
Actions
The five highest-impact actions from Recommendations, ordered by priority. The bars show priority:
Priority
- High priority
- Medium priority
- Low priority
Editors can click Accept to move a suggestion into To do. Click Accepted to undo. Click a title to open the full plan, or View all to see every action.
All chats
Every completed answer in the range. The default order is Most important first: answers where a competitor is recommended and you’re missing come first. A short label under each prompt says why it ranks there.
Views and filters
- Needs attention
- Answers where you lose ground: a competitor named or recommended without you, a competitor ranked above you, a negative answer, and so on.
- Wins
- Answers that recommend you in the top 3.
- All
- Every answer. The Brand menu narrows it to You’re named, You’re recommended, You’re missing or Competitor named, not you. The Sources menu narrows it to Cited, Retrieved only or No sources.
- Group by prompt
- On by default. Shows one row per prompt, with a small icon for each model’s result. Turn it off to see one row per answer, with a Model column.
Why an answer is listed first
Columns
- Chat
- The prompt, its reason label and the start of the answer.
- Your brand
- How the answer treats you, with your rank (#1 = named first). In grouped rows, each model icon gets a badge: green check for recommended, blue check for named, amber ! for warned against, red ✗ for missing.
- Brands named
- Tracked brands in the order the answer names them. +N counts the rest, including brands you don’t track.
- Sources
- “N cited · M retrieved”. Cited sources are shown to the reader; retrieved ones were read but not shown. A blue ring marks your site, a rose ring a competitor’s.
- Scanned
- When the answer was collected.
How an answer treats your brand
- RecommendedSuggested as a good choice
- ComparedWeighed against other options
- MentionedNamed without a verdict
- Warned againstAdvised against
- MissingNot named at all
Click a column header to sort, and click again to reverse. Most important first brings back the default order. Load more adds 50 rows at a time.
Reading an answer
Click any row to open the answer full screen. The browser’s Back button closes it again.
Switch models
When the same prompt was answered by several models, tabs at the top switch between them. A green dot means you were named and a rose dot means you were missing.Read the answer
The prompt appears as a chat bubble, followed by the full answer. Each brand’s quote is highlighted in the colour of its role.Check the evidence
The right panel shows your brand and every competitor the answer names, with rank, sentiment, role, and what the answer praises (+) or criticises (−). Click the quote button to jump to it in the answer. Below that, Cited sources and Retrieved only list every source. When a model only reported a site’s homepage for an article, the link searches that site for the article’s title.Move on
Use ← Previous and Next →, or the ← → keys, to go through the list. Copy answer copies the prompt and the answer. View prompt opens the prompt for editing; closing the editor returns you to the answer.
Before the first results
A new project shows its setup state instead of empty charts:
- Preparing your first scan: setup is complete, and results appear after the first scan finishes.
- Building your answer audit: a scan is running. A progress bar and each model’s status (Queued, Collecting, Processing, Complete…) show how far it has got.
- Allocate prompts / Add prompts / Choose models to start monitoring: something is missing before scans can run. The button takes you to the place to fix it. Only owners and admins can change allocation and models.