PlatformAI visibility
AI visibility tracking
Track how AI assistantsmention your brand
Supasaito asks ChatGPT, Claude, Gemini, Google AI Overviews and Google AI Mode the questions your buyers ask, several times each, and shows whether you are mentioned, in what position and tone, next to which competitors and based on which sources — so you know how AI presents you before buyers ever reach your site.
Free plan: 15 prompts, one scan, 3 samples per prompt · or a one-time extended audit with 75 prompts
How AI visibility tracking works
From buyer questions to trends you can act on
You choose the topics and questions that matter. Supasaito asks the AI models in your plan each question several times, analyses every answer, and turns the results into trends and next steps.
- Your marketCategories and buyer prompts
- Funnel stagesLearn, compare, decide
- Supasaito asks AIAI models in your plan, several samples per prompt
- Every answer analysedMention, position, tone, competitors, sources
- Trends and next stepsWhere you grow and where to act
01Categories and trends
AI visibility by category, tracked scan after scan
Prompts are grouped into categories — your topics, such as product types or use cases. Supasaito tracks visibility, position and sentiment for each category from scan to scan, so you see which topics are growing, which are slipping, and who leads each one.
Categories you control
Group prompts the way your business is organised, so the numbers match your product lines and are easy to report. A prompt has one category or stays Uncategorized.
Trends by category, model and competitor
Switch between categories, AI models and competitors to see where a change comes from. Your publications are marked on the timeline, so you can connect work to results.
Comparable measurements
When prompts are added or moved between scans, the chart marks it, so a changed measurement isn't mistaken for real growth. Prompts that name a brand are kept out.
A page for every category
AI visibility → Categories lists every category with its prompts, visibility, leader and gap to the leader. Drag a row to change the order: the first is the Main category, the default for prompt generation and recommendations. Each category card shows its KPIs, brand ranking, top cited sources and prompts, and generates new prompts for it.
02Prompts and funnel stages
Buyer prompts, grouped by funnel stage
A prompt is a question a buyer might ask an AI assistant. Supasaito writes prompts for your categories from your business context and market research and gives each one a funnel stage — so you measure the questions that lead to sales, and see whether AI names you when buyers learn, compare and decide.
Prompts generated for your categories
Categories with few prompts are filled first, across the whole buying path, so coverage doesn't depend on guesswork. A confirmation shows how many prompts each category gets — at most 20 per category. Keep the prompts, edit them or undo the batch; prompts you delete are never generated again.
Funnel stages
Awareness (“what is…”), comparison and choice (“best…”, “alternatives to…”), decision (pricing, reviews, terms). A gap at the comparison stage means buyers shortlist without you.
Prompts in your buyers' language
Japanese prompts are asked in Japanese. Configured brand aliases and their normalized variants — such as “Aster Modular”, “AsterModular” and its katakana form — are matched to the same brand. Review suggested aliases when a new spelling appears.
03Scans and samples
Several answers per prompt and AI model in every scan
AI answers change from one ask to the next, so in every scan each prompt is asked several times per model. Each answer is a sample. Supasaito averages the samples for every prompt, so one lucky or unlucky answer doesn't mislead your decisions.
3 samples
3 samples
3 samples
3 samples04AI answers
The full AI answers behind every metric
Open any prompt to see a summary for each AI model and the full answers behind it. The actual wording shows how AI describes you, which competitors it puts first and why — the detail you need to decide what to change.
05Competitors in AI answers
Competitors named in each answer, with their position
Every answer records which brands were mentioned and in what position. Across samples you see who AI names next to you, in what order, and which brands you don't track yet. To compare yourself with one competitor in depth, open it in competitor analysis.
Who is named next to you, and in which order
In each answer, every brand is highlighted and numbered by its first mention, so you see your position against the competitors AI lists for this question.
Averaged across samples
For each model you get the average number of competitors, your average position and how often each competitor appears, so one answer doesn't decide.
New brands, added from the answer
A brand you don't track yet appears as a possible competitor. Add it, or mark it as not a competitor, without leaving the prompt.
06Cited sources
The websites AI cites in its answers
Supasaito collects every URL that AI models cite and groups them by domain and source type: your site, competitors, media, Q&A, reviews and more. You learn which websites shape AI answers in your market — the places worth being present on. Domains that also appear in Google results for your Rankings queries show on how many.
Source mix
Owned, competitor, corporate, media, user-generated, Q&A, PR, government, wiki and social. Japanese sources such as note, PR TIMES, ITmedia and Yahoo! Chiebukuro are classified out of the box, so the picture is accurate from the first scan.
Source trends
The share of your own and competitor domains from scan to scan shows whether your presence in the sources AI reads is growing.
Tracked sources
Pin a URL — a new article or a listing — and see when AI starts citing it, so you know a publication has started to work.
07Sentiment analysis
Sentiment of each mention, with the phrase behind it
Every mention of your brand gets a sentiment score, from very negative to very positive, linked to the exact phrase in the answer. You find what AI says that might put buyers off — and where clearer facts about you would correct it.
Five levels, with evidence
Very positive to very negative, each with the sentence that earned it, so you see what to fix, not just a score.
By category and assistant
Where you are spoken of best and which AI model is the most critical, so you know where to start.
Japanese-native tone detection
Tone detection accounts for keigo and the indirect expressions specific to Japanese.
08Ads in ChatGPT answers
Who advertises below the answers to your prompts
ChatGPT can show an ad below its answer. Supasaito keeps the ad apart from the answer text: it is shown as an ad, never counted as a mention, and does not change visibility, sentiment or sources. The Ads in answers page shows how often your prompts come with ads, who advertises and where your competitors pay to appear.
Labelled as ads, not mentions
Each ad is marked as an ad with the advertiser's site. Its statistics live on their own page, so a paid placement never raises anyone's visibility.
Add an advertiser to competitors
An advertiser you don't track yet can be added to your competitors straight from the ad.
By category and prompt
The share of answers with ads, ads by category and the prompts where competitors advertise, over 7, 28 or 90 days. In the prompt list, an Ads column and filter show the same per prompt.
FAQ
AI visibility tracking: common questions
- What is AI visibility?
- AI visibility is how often, and how favourably, AI assistants mention your brand when buyers ask about your market. Supasaito measures it as the share of answers that name you, together with your position, sentiment, the competitors named next to you and the sources the answers rely on.
- How does Supasaito measure AI visibility?
- It sends real buyer questions (prompts) to each AI model several times per scan and analyses every answer. The results are averaged per prompt, category, funnel stage and model.
- Which AI models does Supasaito track?
- ChatGPT, Claude, Gemini, Google AI Overviews and Google AI Mode. The models included depend on your plan.
- How often are prompts scanned?
- Paid plans refresh prompts weekly and allow manual rescans. The Free plan includes one scan of 15 prompts with 3 samples per prompt. A one-time extended audit, paid once without a subscription, scans 75 prompts in 5 categories on ChatGPT, Gemini and AI Overviews instead.
- Why ask each prompt several times?
- AI assistants give different answers to the same question. Several samples per model show how consistently you are mentioned, instead of relying on a single lucky or unlucky answer.
- Does it work for Japanese brands?
- Yes. Prompts are asked in Japanese, configured aliases and their normalized variants, including registered katakana, romaji and short names, are matched as one brand, tone detection accounts for keigo, and Japanese sources are classified out of the box.
- Are ads in ChatGPT answers counted as mentions?
- No. An ad that ChatGPT shows below its answer is stored apart from the answer text, labelled as an ad and kept out of visibility, sentiment and sources. Its statistics are on the Ads in answers page. Ads are detected only in answers captured from the ChatGPT interface.
- Can your team run AI visibility work for us?
- Yes. If you prefer not to run it in-house, our AIO / LLMO consulting program covers measurement, priorities, implementation and reporting.
Related features
Where AI visibility data is used next