What is Mention Rate in AI search?

Mention Rate is the percentage of relevant prompts where an AI engine names your business in its response. If you test 30 prompts across five AI engines and your business is mentioned in 45 of those 150 responses, your Mention Rate is 30%. It is the single most important metric for AI search visibility because it measures whether you are actually appearing when customers ask AI for recommendations.

How often should I check my AI search visibility?

Once a month is the right cadence for most businesses. AI engines update their underlying models and search indexes frequently, so daily or weekly checks produce too much noise. Monthly tracking shows real trends without reacting to short-term variation. Businesses making active changes to their content or schema may benefit from twice-monthly tracking.

Can I track AI search visibility manually?

Technically yes, but not at any useful scale. Manually testing 30 prompts across five AI engines means 150 individual conversations, plus recording, scoring, and tracking changes over time. By the time you finish, the data is already stale. Automated audit platforms run the same prompts in minutes and store results consistently for trend analysis.

Why does AI search visibility differ between platforms?

Each AI engine uses different training data, web sources, and recommendation logic. ChatGPT relies heavily on its training cutoff plus web browsing. Perplexity prioritizes citations from authoritative sources. Gemini integrates Google search signals. Claude is more conservative about specific recommendations. The same business can rank well on one engine and be invisible on another, which is why testing all major platforms matters.

How to Track AI Search Visibility: The Metrics That Actually Matter

Why Tracking AI Visibility Is Not Optional

Every business with a website has spent the last twenty years watching one number: where they rank on Google. SEO tools made that number visible, comparable, and trackable. You knew if you were on page one or page three, and you knew whether things were getting better or worse.

AI search has no equivalent. There is no public ranking. There is no leaderboard. ChatGPT does not show you a list of competitors below your business. When a customer asks Perplexity "Who is the best contractor in Calgary?" you have no idea whether you were named, ignored, or replaced by a competitor — unless you measure it directly.

That measurement gap is the single biggest reason businesses are losing ground in AI search without realizing it. The traffic does not show up in Google Analytics. The lost calls do not show up in your CRM. The customer simply went somewhere else, and you never saw the opportunity.

The One Metric That Matters Most: Mention Rate

Mention Rate is the percentage of relevant prompts where an AI engine actually names your business in its response. It is the foundation of every other AI visibility metric, and the only number that directly answers the question your CFO is going to ask: are we showing up or not?

Core Metric

Mention Rate

The percentage of relevant prompts across AI engines that name your business in the response.

Example: 30 prompts tested across five AI engines = 150 total responses. Your business is named in 45 of them. Mention Rate = 30%.

What good looks like: Strong local businesses see 25 to 50% Mention Rate after AEO work. Below 10% means the AI engines do not have enough signals about your business. Above 60% is exceptional and usually requires sustained content and citation work.

Notice what Mention Rate is not. It is not how many times your URL appears as a citation footnote. It is not website traffic. It is whether the AI, in plain language, recommended you to the user. That is the experience the customer actually sees, and that is what converts.

Secondary Metrics That Add Context

Mention Rate alone tells you whether you are showing up. The next layer of metrics tells you how you are showing up — and whether that representation is helping or hurting your business.

Position in Response

When an AI engine names multiple businesses, where are you in the list? First-named businesses get most of the attention. By the time the AI is on its third or fourth recommendation, the customer has often already decided. Track whether your mentions are leading, middle, or trailing.

Sentiment of Mention

Being mentioned is not automatically good. An AI engine might say "X is reportedly inconsistent on weekends" or "Y has had complaints about pricing." Sentiment analysis on each mention reveals whether the AI is recommending you, describing you neutrally, or actively warning customers away. A negative mention is in some ways worse than no mention at all.

Competitor Comparison

Your Mention Rate in isolation is just a number. Your Mention Rate compared to your three or four direct local competitors is a strategy. If you are at 25% and your top competitor is at 55%, you have a clear gap to close. If you are at 25% and they are at 12%, you have a defensible lead worth protecting. Tracking competitor mentions is what turns AEO from a solo project into a market positioning exercise.

Platform Distribution

You will almost never see uniform performance across all five major AI engines. A business strong on Perplexity might be invisible on ChatGPT. Another might rank well on Gemini but get described incorrectly on Claude. Tracking platform-by-platform reveals which signals are working and which are missing.

Metrics to Ignore

The AI search measurement space is filling up with vanity metrics that sound rigorous but mean nothing for the business. A few worth dismissing immediately:

"AI Visibility Score" with no methodology. Some tools assign a single 0 to 100 score with no explanation of what produces it. If you cannot trace the number back to actual prompts, actual engines, and actual responses, the score is decorative.
Crawl frequency by AI bots. Knowing that GPTBot or PerplexityBot visited your site this week tells you nothing about whether you are being recommended. The bot can read your content and still not surface you when it matters.
Total mentions over time without context. If your "mentions" went from 100 to 200, that could mean you are recommended more, or it could mean someone wrote a Reddit thread arguing about your business. Volume without prompt-level grounding is noise.
Brand search volume. Useful for general marketing, but not a measure of AI search visibility. People search your brand for many reasons, most of which have nothing to do with whether AI is recommending you.

If a metric does not connect back to a specific prompt, a specific AI engine, and a specific response, it is not measuring AI search visibility. It is measuring something else and using AI as a marketing label.

How to Actually Measure These Metrics

The methodology is straightforward in concept but tedious in practice. You define a set of buying-intent prompts your customers might ask. You run each prompt across each AI engine. You record the responses. You tag whether your business was mentioned, where in the response, with what sentiment, and which competitors appeared. Then you do it again next month and compare.

Done manually, this is slow and error-prone. Thirty prompts across five engines is one hundred and fifty individual conversations. Doing it consistently every month for a small business in addition to actually running the business is unrealistic. Doing it in a way that is comparable across months — same prompts, same engines, same scoring rubric — is harder still.

This is the gap automated audit platforms exist to fill. A proper AEO audit runs a fixed set of prompts on a schedule, captures structured results in a database, and produces month-over-month trends without anyone having to type prompts into ChatGPT one at a time. See our complete guide to AEO for the full methodology, or our overview of AEO for local businesses for the broader context.

The Right Cadence: Monthly, Not Daily

Some teams hear "AI search visibility tracking" and assume they need a real-time dashboard. They do not. AI engines update their underlying models and search indexes frequently enough that daily checks produce noisy data without telling you anything actionable. A 2-point Mention Rate change today might be reversed tomorrow with no underlying change to your business.

Monthly tracking is the right cadence for most businesses. It smooths out short-term variation, aligns with the natural pace at which content and citation work compounds, and matches the rhythm at which decisions about content strategy actually get made. Businesses making aggressive AEO changes — publishing weekly content, adding schema, building citations — may benefit from twice-monthly tracking, but daily is overkill for almost everyone.

The exception is when something material changes. A site migration, a major content push, a competitor going quiet — these are moments where an out-of-cycle audit is worth the effort. Otherwise, monthly is enough.

What to Do With the Data

Tracking is only useful if it changes what you do. Three patterns of action come from a well-run measurement program:

Identify gaps. Prompts where you are absent are the most actionable signal in your dataset. Each absent prompt is a customer the AI is sending elsewhere. Address them directly with content that targets that question, schema that clarifies the relevant service, and citations that establish authority on the topic.
Defend wins. Prompts where you appear consistently are working. Resist the urge to leave them alone — they are the most likely to be displaced by competitors who are actively building. Periodic refreshes, additional supporting citations, and review collection on those topics keep the position durable.
Pivot when sentiment is negative. If the AI is naming you but describing you in unflattering terms, that is a brand and content problem, not a visibility problem. Look at what content the AI is drawing from, address the underlying issues — outdated information, unflattering reviews, missing context — and re-test in the next cycle.

Common Mistakes Businesses Make When Tracking

A few patterns show up repeatedly in businesses that try to track AI visibility on their own:

Changing the prompts every month. If the prompts change, the comparison breaks. Pick a stable set of buying-intent prompts and only revise them when the underlying business genuinely changes.
Testing only one platform. ChatGPT is the loudest AI engine, but it is not the only one. Customers using Perplexity, Gemini, Copilot, and Claude all matter, and each can produce a different picture of your visibility.
Confusing brand mentions with recommendations. An AI saying "There is a business called Yours" is not the same as an AI recommending you. Score for actual recommendations, not for any appearance of the name.
Quitting after one month of poor results. AEO is a compounding discipline. The signals that drive AI mentions take time to accumulate. Three months of consistent measurement and intervention is the minimum window to see meaningful trend changes.

The Bottom Line

What you measure is what you can improve. AI search visibility is no different from any other business discipline in that respect — the businesses that track Mention Rate, sentiment, position, and competitor performance every month will outperform the ones that do not, simply because they are operating with information instead of guessing.

The investment is small. The payoff is durable. Start measuring this month and you will be twelve months ahead of competitors who are still hoping AI will find them on its own.