To measure AI search visibility, you need to track three things: how often your brand is mentioned for the prompts your buyers ask, how often your pages are cited as sources, and how your presence compares with competitors across those answers. This article explains how to measure each layer, run a manual visibility audit with a spreadsheet, choose between tracking tools, and decide how often to measure progress.
Why you can’t measure AI search visibility like keyword rankings
Traditional SEO gave marketers a relatively stable object to measure. We used to deal with a search results page with positions that we checked and tracked over time.
But AI answers don’t behave like this. The same prompt can produce different responses across runs. Outputs can also vary by user context and model version. ChatGPT, Gemini, Perplexity, AI Overviews and other engines do not all retrieve or cite information in the same way.
So we no longer have a single “position three” to monitor. Instead, we have a sampling problem that we need to solve continuously:
- You define a meaningful set of prompts
- Then you run them repeatedly across the engines
- Based on those runs, you can measure how often your brand appears, and which competitors show up alongside it
So when you measure AI search visibility, the metric that matters is frequency across a representative prompt set over time.
The three metrics that matter: brand mentions, citations, share of voice
Brand mentions
These tell you how often an AI engine names your company across the prompt set you care about. Expressed as a percentage of prompt runs, this is the clearest indicator of basic visibility. The limitation is context: a mention does not automatically mean recommendation, endorsement, or buying intent.
Citations
Your citation rate tells you how often your own URLs appear as linked sources in AI-generated answers. This helps you see which pages show up as evidence, and on which engines. But citations need interpretation. An engine may mention your brand while citing a third-party article, review site, or media source instead of your own website.
AI share of voice
AI share of voice compares your visibility with named competitors across the same prompt set. It is useful for understanding whether your presence is growing relative to the market rather than in isolation. The main requirement is consistency: change the competitor set too often and your trend line stops meaning much.
| Metric | What it measures | How it’s expressed | Watch out for |
|---|---|---|---|
| Brand mentions | How often engines name you for your target prompts | % of prompt runs that mention the brand | Mentions without recommendation intent |
| Citations | How often your URLs appear as linked sources | Citation count / cited pages per engine | Engines may cite third-party pages about you |
| AI share of voice | Your mentions vs named competitors on the same prompts | % share across the prompt set | Changing the competitor set breaks trend continuity |
These metrics are all incomplete on their own. It’s only when you combine them that you get a clear picture. That picture should show you:
- Whether you appear in AI search, or you don’t
- How AI engines are using your content
- Your presence in relation to your competitors
How to run a manual AI visibility audit
You do not need a specialist platform to measure AI search visibility and establish a useful baseline. Grab a spreadsheet, then try these steps.
- Build a set of 20–30 buyer questions. Use questions that reflect real buying journeys rather than prompts designed to mention your brand. Mine sales calls, support tickets, search data, and Google’s People Also Ask results for recurring themes.
- Run the same prompts across multiple AI engines. Start with ChatGPT, Gemini, and Perplexity. Where possible, use fresh sessions and minimize account history or personalization so your tests are easier to compare.
- Record what appears. For every run, capture whether your brand showed up, whether your website got a citation, which URL appeared, and which competitors the model named alongside you.
- Run each prompt three times. AI answers can vary between identical requests. Repetition helps reduce the influence of a single unusual response.
- Calculate your baseline metrics. From the results, calculate brand mention rate, citation frequency, and share of voice against a fixed competitor set. A simple spreadsheet is sufficient.
- Repeat the identical audit monthly. Keep the prompts, engines, competitor set, and recording method as stable as possible. That consistency is what turns a collection of AI answers into trend data.
You can achieve a lot with relatively little here. A consistent set of 25 meaningful prompts tracked month after month is more useful than 200 prompts that keep changing.
Tools for tracking AI visibility
Tools that measure AI search visibility mostly automate the same sampling process described above. Their real value lies in scale. Instead of manually running prompts, recording mentions, and calculating share of voice, the software does it continuously.
Dedicated AI visibility platforms such as Profound, Otterly.ai, and Peec AI are built specifically for this job. They track prompts across multiple engines, monitor brand mentions and citations, and benchmark competitors. Pricing varies widely: lighter tools can start relatively cheaply, while enterprise platforms tend to move quickly into custom pricing.
SEO suites are adding AI visibility as another reporting layer. Semrush now offers AI Visibility tracking from $99 per month per domain, including 25 custom prompts, while Ahrefs offers custom prompt tracking from $50 per month and its broader Brand Radar AI product from $199 per month.
DIY tracking is also viable. Teams with technical resources can use APIs and scripts to run a fixed prompt set, store responses, and calculate mentions, citations, and competitor share of voice themselves. The software cost may be lower, but the setup, maintenance, and normalization work moves in-house.
At Inspired Marketing, we also run our own internal tracking across client-specific prompt sets so we can monitor visibility consistently rather than relying on one-off checks.
How often should you measure AI search visibility?
For most B2B brands, monthly measurement is enough. AI visibility moves too slowly and too noisily for daily checking to be useful. Weekly tracking makes more sense during an active campaign, after a major model release, or when you have shipped a significant batch of new content and want to watch for early movement.
Expect changes to take time. In our experience, meaningful shifts often start to show 6–12 weeks after new content goes live, depending on the topic, authority of the domain, and how quickly different engines refresh their source sets.
What is “good” in AI search visibility?
There is no reliable industry benchmark for “good” AI visibility yet, and we would be wary of anyone pretending otherwise. For a newer or previously invisible brand, appearing in 10–20% of relevant category prompts within a quarter can represent meaningful progress. The key thing to watch is the overall trend: are mentions, citations, and share of voice improving across the same prompt set over time?
What to do with the data
The value of AI visibility tracking is in what it tells you to change.
Low mentions and low citations usually point to a foundational content gap. The brand may simply lack enough clear, authoritative material for AI engines to retrieve and use. That is where an [AEO readiness checklist] can help identify missing pieces.
Mentions without citations tell a different story. Your brand is known, but other websites are doing the explaining. That is a strong signal to invest in original data, digital PR, expert commentary, and third-party coverage that gives engines better material to work with.
Citations without mentions mean your content is being used as evidence, but your brand is not getting explicit visibility from it. Semrush found 62% of citations appeared without an explicit brand mention. The opportunity is to make authorship, proprietary data, and brand attribution harder to separate from the information itself.
Losing share of voice on one engine is a cue to look at the sources that engine is using. Which domains appear repeatedly? Which competitors are being cited, and from where?
At Inspired Marketing, our AEO work includes this kind of visibility tracking so we can connect measurement back to content and authority-building decisions. Talk to us about a baseline audit.
