TL;DR
AI search visibility is how often, and how well, your brand appears in answers from ChatGPT, Perplexity, Gemini, Copilot, Google AI Overviews and AI Mode. It has two parts. A mention is when the AI names your brand. A citation is when it links to your website as a source. You need to track both.
There is no single number for it, and no tool gives you the full picture. A reliable measurement uses four sources together: a fixed set of buyer prompts you test regularly, Google Search Console’s new generative AI report for AI Overviews and AI Mode, AI referral traffic in GA4, and a tracking tool if your budget allows one.
Start with a free manual baseline of 20 to 50 prompts. Measure mention rate, citation rate, share of voice against two or three competitors, and how accurately AI describes you. Repeat it monthly. Then turn the gaps into work: pages to create, pages to update and third-party sites to be mentioned on.
What Is AI Search Visibility?
For twenty years, search visibility meant one thing: where your pages ranked on Google. That definition no longer covers how B2B buyers research. A head of marketing shortlisting agencies, or a CTO comparing software, now asks ChatGPT or Perplexity for options, reads a Google AI Overview, and often visits only the two or three vendors the answer names.
AI search visibility, also called AI brand visibility or LLM visibility, measures your presence in those answers. It covers three questions:
- Are you mentioned? Does the AI name your brand when someone asks a question in your category?
- Are you cited? Does it link to your website as a source, which can send traffic?
- How are you described? Is the description accurate, positive and in line with how you position yourself?
Don’t treat mentions and citations as the same thing. An AI tool can recommend your company by name while citing a review site, a Reddit thread or a competitor’s listicle as its source. It can also cite one of your blog posts for a definition without ever recommending you. Research from AirOps suggests that about 85% of brand mentions in AI search come from third-party pages rather than the brand’s own site, so a citation-only view will miss most of what is said about you.
If you are new to the wider shift, our article on GEO vs SEO explains how optimising for AI answers differs from traditional SEO. This article focuses on one part of that work: measurement.
Why AI Visibility Is Harder to Measure Than Rankings
A Google ranking is fairly stable. If you rank fifth today, you will probably rank somewhere near fifth tomorrow, and Search Console tells you which queries brought clicks. AI answers do not work that way.
Ask ChatGPT the same question three times and you may get three different lists of vendors. AirOps’ 2026 research found that only about 30% of brands stayed visible from one AI answer to the next, and around 20% stayed visible across five consecutive runs. Answers also change with the user’s location, conversation history, model version and whether the tool searched the web before answering.
There is a data problem too. Google shows you queries and clicks for organic search. ChatGPT and Perplexity show you nothing at all. You cannot see which prompts people typed, how often your brand appeared or how many people read the answer without clicking.
So treat AI visibility as a sample, not a census. You are estimating your presence from a representative set of prompts, run consistently over time. The trend tells you more than any single result.
The Metrics That Actually Matter
Every tool names these differently. Underneath, there are eight numbers, and you do not need all of them on day one.
| Metric | What it measures | How to calculate it |
|---|---|---|
| Mention rate | How often AI names your brand | Answers that mention you ÷ total prompts tested |
| Citation rate | How often AI links to your site | Answers citing your domain ÷ total prompts tested |
| Share of voice | Your presence relative to competitors | Your mentions ÷ total mentions of you and named competitors |
| Prominence | Where you appear in the answer | First mention, in the list, or only in passing |
| Sentiment and accuracy | How AI describes you | Manual review: positive, neutral, negative; correct or incorrect |
| Cited pages | Which of your URLs AI relies on | List of your pages that appear as sources |
| AI referral traffic | Visits from AI tools to your site | GA4 sessions from AI referrers |
| AI Overview impressions | How often your pages appear in Google’s AI features | Search Console generative AI report |
If you only track two, track mention rate and share of voice for unbranded prompts. Those tell you whether AI recommends you to buyers who have not heard of you yet, which is where most new pipeline comes from. Branded prompts (“What is EverClif?”) matter for accuracy, but they do not show discovery.
How to Measure AI Search Visibility, Step by Step
The method below works for a B2B company of any size. The first two steps are free and take an afternoon. The later steps add data you already have access to, and a paid tool only becomes necessary once you want to track hundreds of prompts.
Step 1: Build a Prompt Set Your Buyers Would Use
Everything depends on the prompts you test. Choose the wrong ones and you will measure visibility for questions nobody asks.
Build your list from real buyer language, not from your keyword list. The best sources are sales call notes, demo request forms, support tickets, Search Console queries, and the questions people ask in Reddit and LinkedIn threads about your category. Then sort the prompts into four groups:
| Prompt type | Example for a B2B SaaS marketing agency | What it shows |
|---|---|---|
| Category | “Best B2B SaaS marketing agencies in the US” | Whether you are on the shortlist |
| Problem | “How do I get more demo requests from organic search?” | Whether your expertise is cited |
| Comparison | “[Your brand] vs [competitor] for SaaS SEO” | How you are positioned against rivals |
| Branded | “What does [your brand] do and what does it cost?” | Whether AI describes you accurately |
Twenty to fifty prompts is enough for a first baseline. Weight the list towards category and problem prompts, because that is where buyers discover vendors. Write the list down and keep it fixed. If you change the prompts every month, you cannot compare results.
Step 2: Run a Manual LLM Visibility Check
Before paying for any LLM visibility checker, run the prompts yourself. It is slow, but it teaches you more about how AI sees your category than any dashboard.
Test each prompt in ChatGPT, Perplexity, Gemini and Google (for AI Overviews and AI Mode). Use a logged-out or private session where possible, so your own history does not skew the answer. Because answers vary, run each important prompt at least three times and record how many of those runs mention you.
Log the results in a simple spreadsheet with these columns: prompt, prompt type, platform, date, mentioned (yes/no), cited (yes/no), position in the answer, competitors named, sources cited, and notes on accuracy or tone. The “sources cited” column is easy to skip and should not be. It shows you which publications, review sites and forums AI trusts in your category, and those are the places you need to be mentioned.
Step 3: Track Google AI Overviews and AI Mode
Google is the one AI platform that gives you first-party data. In June 2026, Google introduced Search Generative AI performance reports in Search Console. They show how often your pages appeared in AI Overviews and AI Mode, which URLs appeared, and a breakdown by country, device and date. The data is a separate view of numbers that were already included in your overall Performance report.
The reports have clear limits. They show impressions, but not clicks, click-through rate, average position or the queries that triggered the appearance. Data starts from 18 May 2026, and Google rolled the reports out gradually, so check whether they appear in your property yet. Use them for trends, such as whether AI impressions rise after you update a page, rather than as an absolute measure. For tracking AI brand visibility in Google specifically, this is the most reliable free source you have.
For query-level AI Overviews tracking, you need a third-party tool. Semrush, Ahrefs and SE Ranking all show which tracked keywords trigger an AI Overview and whether your domain is cited in it. For a small keyword set, you can also check by hand: search the query, expand the AI Overview and note the cited sources.
Step 4: Measure AI Referral Traffic in GA4
Mentions and citations are leading indicators. Traffic and leads are what your leadership team will ask about. GA4 does not classify AI traffic by default, so visits from ChatGPT or Perplexity land in Referral, Direct or Unassigned unless you set it up.
The simplest setup takes about ten minutes:
- In GA4, go to Admin → Data display → Channel groups and create a new custom channel group.
- Add a channel called “AI Search” where the session source matches this regex:
chatgpt\.com|chat\.openai\.com|perplexity|gemini\.google|copilot\.microsoft|claude\.ai|meta\.ai|deepseek - Move the new channel above “Referral” in the list so it is matched first, then save.
- Build an Exploration with the AI Search channel, landing page and key events (demo requests, contact forms) to see which pages AI sends people to and whether they convert. The Google Analytics MCP server can help you query the same landing-page and key-event data during recurring reviews.
Two caveats. First, not all AI traffic carries a referrer. Some apps and in-app browsers strip it, so a share of AI-driven visits will appear as Direct. Second, many buyers read an AI answer, remember your name and search for you on Google later. That visit shows up as branded organic search, not AI. A “How did you hear about us?” field on your demo form, with ChatGPT and Perplexity as options, fills part of that gap and is worth adding today.
Step 5: Benchmark Share of Voice Against Competitors
Your mention rate means little on its own. Being mentioned in 20% of answers is strong in a crowded category and weak in a niche one. Share of voice puts the number in context.
Pick two or three direct competitors, the ones you lose deals to, and count their mentions across the same prompt set. Your share of voice is your mentions divided by the total mentions for you and those competitors. Then look at the pattern behind the number. Which prompts do competitors win that you do not? Which sources do the answers cite when they recommend them? That list becomes your content and PR plan.
Step 6: Set a Reporting Cadence
Monthly is the right rhythm for most B2B teams. Weekly checks mostly capture random variation in AI answers, and quarterly checks are too slow to show whether your changes worked.
A monthly AI visibility report needs only five lines: mention rate, citation rate, share of voice, AI Overview impressions from Search Console and AI referral sessions with conversions from GA4. Add a short note on any inaccurate descriptions you found and what you are doing about them. Keep the prompt set the same across months, and add new prompts as a separate group so they do not distort the trend.
What We Found When We Measured EverClif
We ran this check on our own site before writing this article. The numbers were not flattering.
Using Ahrefs’ AI citation data in October 2026, everclif.com had 6 citations across 3 pages, all in Microsoft Copilot. We had zero citations in ChatGPT, Perplexity, Gemini, Google AI Overviews and AI Mode. For comparison, an established B2B SaaS marketing agency we compete with had 414 citations across 98 pages over the same platforms.
One detail stood out. The only platform citing us was Copilot, which draws on Bing’s index, while ChatGPT, Perplexity and Google ignored us entirely. Each AI tool has its own sources, so a brand can be visible in one and invisible in another. That is why you test every platform separately.
The bigger gap is footprint. The competitor has years of published pages, backlinks and third-party mentions, and AI tools draw on that history. Our gap tells us exactly what to do next: publish pages that answer the questions buyers ask AI, get those pages indexed and linked, and earn mentions on the sites AI already cites for our category. We will update this section as the numbers move.
Best AI Visibility Tools: Free vs Paid
Once you have a manual baseline, a tool saves time and lets you track far more prompts. The market has grown quickly, and most tools fall into three groups.
| Option | What it covers | Best for |
|---|---|---|
| Manual testing + spreadsheet | Any platform you can access; full control of prompts | First baseline, small budgets |
| Google Search Console + GA4 | AI Overviews and AI Mode impressions; AI referral traffic | Every team; free first-party data |
| Free AI visibility checkers (e.g. Ahrefs’ free checker) | Quick snapshot of brand presence | A fast first look |
| Semrush AI Visibility Toolkit | Mentions, citations, share of voice, prompt research and sentiment across major AI platforms | Teams already using Semrush; included in its SEO + AI Search plans |
| Ahrefs Brand Radar | Mentions, citations, impressions and share of voice across AI platforms, backed by a large prompt database | Teams already using Ahrefs; sold as an add-on |
| SE Ranking AI search tracking | AI Overviews and AI tool tracking at a lower price | Smaller teams and agencies |
| Dedicated AI visibility platforms (e.g. Profound, Peec AI, Otterly.AI, Scrunch) | Custom prompt tracking, multiple models, detailed reporting | Teams making AI search a core channel |
Our advice is to use the tool you already pay for before buying another. If you use Semrush or Ahrefs for SEO, their AI visibility features will cover most B2B needs. Our Semrush vs Ahrefs comparison covers how their AI visibility features differ. Dedicated platforms make sense when you need to track hundreds of custom prompts across many models, or report AI visibility to a board as a headline metric.
Whatever the tool, check which prompts it tests. Ahrefs Brand Radar, for example, reports on its own large prompt database by default, and custom prompts need a separate report set up rather than your buyers’ questions. Those numbers are useful for comparison, but they are not the same as your manual baseline.
What to Do With the Numbers
A number on a dashboard changes nothing. We sort every finding into one of three types of work.
If ChatGPT is the gap in your baseline, our ChatGPT SEO guide turns those findings into a plan for crawl access, source content and citations.
Fix. Correct anything AI says about you that is wrong: outdated pricing, a service you no longer offer, a wrong location. Update your own pages first, then the third-party profiles AI cites, such as directories, review sites and partner listings.
Build. For each prompt where competitors appear and you do not, check whether you have a page that answers the question directly. Often you do not. Create it, structure it with clear headings and short, direct answers, and keep it up to date. Our guide on how to optimise for AI search covers the page-level changes in detail.
Influence. Look at the sources AI cites when it recommends your competitors. Those review sites, industry publications, listicles and forums are where you need to be mentioned. This is the slowest work and often the most effective, because AI tools lean heavily on third-party sources.
Common Measurement Mistakes
- Tracking only branded prompts. AI will almost always mention you when asked about you by name. That tells you nothing about discovery.
- Trusting a single run. One answer is an anecdote. Run important prompts several times and track the rate, not the result.
- Changing the prompt set every month. You lose the ability to see a trend. Keep a fixed core set and add new prompts separately.
- Counting citations and ignoring mentions. Most mentions come from third-party sources. A citation-only report misses them.
- Expecting AI traffic to look like SEO traffic. AI referral volumes are usually small. Judge them by conversion rate and lead quality, and remember that a lot of AI influence shows up later as branded search.
- Measuring without acting. A dashboard nobody uses is a cost, not an asset. Every monthly report should end with two or three actions.
Final Thoughts
You will not measure AI search visibility as precisely as Google rankings. Perhaps not ever. Measure it anyway. Your buyers are already asking AI tools for recommendations, and the brands those tools name are the ones that make the shortlist.
Start small. Write down 30 prompts your buyers would ask, test them this week, set up the AI channel in GA4 and check the generative AI report in Search Console. In a month, run the same prompts again. You will have a baseline, a trend and a clear list of pages and mentions to work on.
You cannot improve what you have never checked. Ask the AI about your category before your buyers do.