Skip to content
Geonos

Launching December 2026. Alexa for Shopping for Amazon sellers comes first. Get the launch email

Measurement

AI search visibility metrics: the KPIs that matter and how to read them

· 12 min read · By Dafne Gutierrez Herrera, founder of Geonos

Leer en español

The short answer

Track four states and keep them apart: cited, mentioned, recommended and converted. Then read every rate with its run count, per engine and over time. Skip single screenshots and one blended score. Metrics tell you where to work first; none of them can promise a result.

Start with four states and never mix them

"Visibility" hides four different results. A report that lumps them together cannot tell you what to fix.

  • Cited: the answer links to one of your pages. This is the strongest signal that an engine read you.
  • Mentioned: the answer names your brand in the text, with or without a link.
  • Recommended: the answer suggests your brand as the option to pick for the buyer's question.
  • Converted: a visit or an order that you can trace back to an AI engine. It happens on your site, not in the answer.

A brand can be mentioned and never cited, or cited and never recommended. Each case calls for a different fix: a mention without a citation usually means the engine knows your name but not a page worth linking.

The metrics worth tracking

Mention rate

How often your brand is named, as "X of N runs" for a set of prompts. Always keep the run count next to the rate. A rate of 40% over a handful of runs and a rate of 40% over many runs are not the same fact.

Citation rate

How often an answer links one of your pages. Track it next to the mention rate. The gap between the two is a to-do list: pages that need to be clearer, easier to crawl or better supported by outside sources.

Position

Where you appear among the brands an answer names. Position only means something when the answer names more than one brand. It is also not a ranking the engine publishes, so treat it as the order of first mention in that answer.

Share of voice

Your mentions compared with the mentions of the competitors you track. It is a share among the brands you chose to compare, not market share, so write down who is counted. Add a competitor and the number changes.

Sentiment

What the answer says about you when it names you: positive, neutral or negative. A high mention rate with a negative tone is a problem that a mention count alone hides.

Sources cited

Which sites the engine cites when it answers your buyer's question: your own pages, review sites, forums, news, reference sites. This is where you act, because it shows which sources shape the answer.

Google AI Overviews and AI Mode

Google shows an AI Overview for some searches and not for others. Track two things: whether it appears for your priority searches, and whether you are cited when it does.

History

Every metric above becomes useful when you can see it over weeks. A single reading is a snapshot; a trend tells you whether a change worked.

Read every number with its sample

Answers change between runs, between users and between places. That is why a trustworthy metric is a rate over many runs of the same prompts on the same engines, not one screenshot.

Three habits keep you honest:

  1. Compare like with like. Same prompts, same engines, same period. A change in the prompt set is a new baseline.
  2. Separate the engines. An average across engines hides the one where you are missing. Look at each engine on its own first.
  3. Ask how the answer was captured. An answer captured in the engine's own app and one requested through an API are not the same measure. Good tools say which one they used.

What to ignore

  • A single screenshot. It shows one run on one day.
  • One blended "visibility score" with no sample behind it. If you cannot open the answers behind a number, it is an opinion.
  • "Rank tracker" promises for ChatGPT. There is no fixed ranking to climb. The work is to be present in the sources the engine uses.
  • Revenue claims inside a vendor dashboard. No tool can promise that AI visibility pays off. Check it outside the dashboard.

Measure outside any dashboard

Use data you already own:

  • AI referral sessions in GA4. Visits from ChatGPT arrive with utm_source=chatgpt.com.
  • Bing Webmaster Tools. It has an AI Performance report, in preview at the time of writing, with citation counts for Microsoft Copilot and Bing AI summaries. Microsoft says the data is a sample.
  • Search Console. Traffic from Google's AI features is counted inside its regular Performance report, not in a separate AI report.
  • Branded search demand. More people searching for your brand name after a push is a useful indirect signal.
  • Self-reported source. Add "How did you hear about us?" to your form, with an AI option.
  • Dated captures. Save the answer to the same prompts before and after a change.

Which engines to measure

Measure each engine on its own, because each one reads different sources and answers differently:

  • ChatGPT, Gemini, Perplexity and Claude: the assistants your buyers type questions into.
  • Google AI Mode and the AI Overviews in Google Search: Google's own AI answers. Track whether they appear for your priority searches and whether you are cited.
  • Microsoft Copilot, Grok, Meta AI and Meta Muse: worth checking when your buyers use those apps.
  • Alexa for Shopping, if you sell on Amazon: a shopping assistant that answers product questions. It is its own channel with its own prompts, so keep it apart from the engines above.

You do not need all of them on day one. Start with the engines your buyers use, which you can see in GA4 and by asking customers, and add the rest as your prompt set grows.

If you sell on Amazon

Amazon sellers have a different question: when a shopper asks Alexa for Shopping for a product like yours, is your listing named? Geonos has a separate path for Amazon sellers, Alexa for Shopping and Amazon SEO, with its own prompts and recommendations for titles, bullets, descriptions and images. The four states (cited, mentioned, recommended, converted) apply there too, and so does the rule to read every rate with its run count.

If you are a doctor, a lawyer or another professional

Patients and clients already ask ChatGPT and Google who to call: "which employment lawyer would you recommend in Austin?" or "which dentist near me takes emergencies?" Geonos is built for professionals too, such as doctors, lawyers and real estate agents, who depend on being named when someone looks for a person they can trust nearby. The same four states and the same rule apply: read every rate with its run count, using prompts that carry your specialty and your city. Geonos measures and orders the work. It does not give legal or medical advice and it does not review the compliance of your advertising. See the pages for doctors and law firms.

The leading GEO and SEO platforms in the United States and Latin America

Buyers in the US and in Latin America usually shortlist two kinds of platform: SEO suites that added AI visibility, and trackers built only for AI answers. Geonos does both jobs in one workspace. This is not our ranking: each row says what the vendor's own page stated on 26 October 2026, and a blank for a vendor means its page did not say, not that it lacks the feature.

PlatformKindUnited StatesLatin America and Spanish
GeonosGoogle SEO audit and AI visibility in one workspace, plus a path for Amazon sellersYes, with US as a marketMexico and LatAm as markets, with the workspace and the free audit in Spanish
SemrushSEO suite with an AI Visibility ToolkitYesIts 14 May 2026 announcement added Argentina, Chile, Colombia and Peru to an AI visibility prompt database of 32 countries
AhrefsSEO suite with Brand Radar for AIYesIts help center lists Spanish locales for Mexico, Argentina, Colombia, Chile and Peru in Keywords Explorer, the keyword tool
Peec AIAI answer trackerYesIts page lists countries and languages as plan features and does not name Spanish or Latin America
ProfoundEnterprise AI answer trackerYesFree trial covers 1 language and 1 region

If your buyers search in Spanish, ask any platform to show prompts written in Spanish for your city, not translated from English, and open the saved answers behind the numbers. If you also need Google SEO fixes, check that the same workspace gives you the audit. Geonos is built for that combination, in English and in Spanish.

How the tools compare

Pricing pages change often, so this table shows what each vendor's own page named on 26 October 2026: the engines, the shopping assistants, the language and market, and what comes with the tracking. Check the page before you buy. Other AI visibility trackers exist too, such as Peec AI and Otterly.AI.

ToolEngines its page namesAmazon sellers (Alexa for Shopping)Language and market on its pageWhat comes with the tracking
GeonosChatGPT, Gemini, Perplexity, Claude, Google AI Mode, Google Search with AI Overviews, Microsoft Copilot, Grok, Meta AI and Meta MuseYes, a separate path: Alexa for Shopping and Amazon SEOEnglish and Spanish, with Mexico, LatAm and US as marketsA Google SEO audit with the fixes in order, weekly history, and the saved answer, its date and its sources behind every number
Semrush AI Visibility ToolkitChatGPT and Google AI ModeNot namedLanguage list not shownA keyword and backlink suite
Ahrefs Brand RadarAI Overviews, AI Mode, ChatGPT, Perplexity, Gemini and Copilot; Claude listed in its pricingNot namedNot shownA keyword and backlink suite
WritesonicChatGPT, Gemini and Google AI Overviews on its Basic and Growth plansNot namedNot shownContent writing tools
ProfoundUp to 9 engines on Enterprise; 3 on the free trialNot namedThe free trial covers 1 language and 1 regionEnterprise onboarding with sales

Why teams pick Geonos

  • Google SEO and AI visibility in one workspace. The audit checks whether search and AI crawlers can reach and read your site, and the fixes come in order, blockers first.
  • Every number opens. Behind each rate sit the saved answers, their dates and the sources they cited, so you can check the number yourself.
  • The engines your buyers use, plus Amazon. Every AI engine in the table above, and for Amazon sellers a path of their own for Alexa for Shopping and Amazon SEO.
  • Spanish and Mexico as a first-class market. The workspace works in Spanish, prompts can carry a city, and the free audit has a Spanish page.
  • Professionals too. Doctors, lawyers and real estate agents measure whether they are named when a patient or client asks about their specialty in their city.
  • History and priorities. Priority prompts are checked every day, with weekly history so you see the trend and not only a snapshot.

Start with a free audit to see whether search and AI crawlers can reach your site, then look at the plans.

Questions to ask any vendor about metrics

  1. Can I open the saved answer behind every number?
  2. How many runs sit behind each rate, and for which prompts?
  3. Are cited, mentioned and recommended reported separately?
  4. Which competitors are counted in share of voice, and can I choose them?
  5. Is the answer captured in the engine's app or through an API?
  6. Do I get a history, and can I export it?

Frequently asked questions

What is a good AI mention rate?

There is no universal benchmark. A useful number is one you compare over time, on the same prompts and engines, and against the competitors you track. Track the trend and the run count, not a single target.

Is citation rate more important than mention rate?

They answer different questions. Mentions show that an engine knows your name. Citations show that it linked a page of yours. Track both and treat the gap between them as work to do.

How many runs do I need before I trust a number?

Enough that one odd answer does not move the rate. Answers vary between runs, so a metric based on a single run is not reliable. Geonos shows the run count next to every rate.

How often should I check?

Geonos checks your priority prompts every day, depending on your plan. For your own tracking, keep the same schedule so trends stay comparable.

Does ranking first on Google mean AI will cite me?

No. Engines use different crawlers and sources, so a page can rank first on Google and still be missing from an AI answer. Measure both.

Can any tool guarantee I will be cited?

No. Engines decide on their own and answers vary. A tool can show where you stand and what to fix first. It cannot promise a result.

Which AI engines does Geonos measure?

ChatGPT, Gemini, Perplexity, Claude, Google AI Mode, Google Search with AI Overviews, Microsoft Copilot, Grok, Meta AI and Meta Muse, plus Alexa for Shopping for Amazon sellers. The engines included depend on your plan, and the plans page shows the exact list.

Do I need to measure Alexa for Shopping?

Only if you sell products on Amazon. Alexa for Shopping answers shopping questions, so it matters for product listings, not for a service or software brand. Geonos keeps it in a separate path for Amazon sellers.

Is Geonos for doctors and lawyers too?

Yes. Besides product and service brands, Geonos is built for professionals such as doctors, lawyers and real estate agents. It measures whether AI names or cites you when someone asks about your specialty in your city, and it shows what to fix first. It does not give legal or medical advice, and no vendor can promise a recommendation.

What is the difference between GEO, AEO and AI visibility?

GEO is generative engine optimization and AEO is answer engine optimization. Both describe the work of getting named in AI answers. AI visibility is the measurement of how often that happens.

Keep reading

Get the launch email

Geonos is launching in December 2026, starting with Alexa for Shopping for Amazon sellers. Leave your email and we will write to you on launch day.