How do I track whether Google AI Overviews mention my brand?

To track whether Google AI Overviews mention your brand, run a fixed list of 20–50 customer questions in Google on a set schedule, record whether an AI Overview appears, whether your business is named or linked, and which competitors and sources are cited, then repeat the same check every week or month and compare the results. Found in AI automates this process for local businesses by tracking how AI tools answer customer questions and showing where your brand is missing.

The rest of this guide explains how to build that tracking process, what to record, how to read the results, and how to close the gaps you find. It is written for local businesses and service providers: plumbers, HVAC companies, roofers, dentists, med spas, chiropractors, law firms, accountants, financial advisors, restaurants, auto repair shops, and fitness studios.

Why tracking AI Overviews is different from tracking rankings

Traditional rank tracking answers one question: what position does my page hold for this keyword? AI Overviews change the question. Google generates a summary at the top of the results page, pulls from several sources, and may name businesses, link to pages, or mention no one at all. A page can rank third in the standard results and never appear in the summary. A business can also be named in the summary without ranking on page one for the term.

That means position numbers alone no longer describe your visibility. You need to measure four separate things:

• Presence: does an AI Overview appear for the query at all?
• Mention: is your business name stated in the summary text?
• Citation: is your website linked as a source?
• Context: what does the summary say about you, and is it accurate?

A business that tracks only rankings misses all four. A business that tracks these signals sees exactly where customers are getting answers and whether those answers include it.

Step 1: Build a question list that matches how customers ask

Tracking starts with the right queries. AI Overviews appear most often on question-style and research-style searches, so your list should reflect how real customers phrase problems, not just short keywords.

Build your list from these sources:

• Phone calls and emails: write down the exact questions customers ask before they book.
• Google autocomplete and "People also ask": type your service and city and note the suggestions.
• Review text: customers describe their problems in their own words.
• Your own service pages: each service you sell should generate at least three questions.

Sort the questions into groups so the results are easy to read later:

• Problem questions: "why is my furnace blowing cold air," "what causes a root canal to hurt after treatment."
• Cost questions: "how much does a roof replacement cost in [city]."
• Comparison and selection questions: "how do I choose a family law attorney," "what should I look for in an auto repair shop."
• Local intent questions: "best med spa near me," "emergency plumber in [city]."
• Brand questions: "is [your business name] good," "[your business name] reviews."

Aim for 20–50 questions. Fewer than 20 gives you noisy data. More than 50 becomes hard to maintain by hand. Once the list is set, do not change it casually. Tracking only works when you compare the same questions over time. Add new questions in a separate batch so your baseline stays clean.

Step 2: Run clean manual checks

Manual checking is the free starting point, and it teaches you how AI Overviews behave. Results vary by location, device, search history, and signed-in status, so control those variables.

1. Use a private or incognito window so personalization has less influence.
2. Stay signed out of your Google account when possible.
3. Search from the area you serve. Local queries change by location. Use a consistent location for every check, and add the city name to the query where customers would.
4. Use the same device type each time. Mobile and desktop results can differ.
5. Check on a fixed schedule. Weekly suits fast-moving markets; monthly suits most local businesses.

For every query, log the result in a spreadsheet with these columns:

• Date of check
• Query text
• Query group (problem, cost, comparison, local, brand)
• AI Overview shown (yes or no)
• Business named in text (yes or no)
• Website linked as a source (yes or no)
• Other businesses or sites named
• Notes on accuracy (hours, services, location, claims)

Take a screenshot of each Overview. AI answers change, and a screenshot is the only record of what a customer saw on a given day.

Step 3: Expect variation and test each query more than once

AI Overviews are generated, not fixed. The same query can produce slightly different wording, different sources, or no summary at all from one day to the next. A single check does not prove you are present or absent.

Handle this with a simple rule: run each query at least three times across different days within a reporting period, and count a mention as real when it appears in at least two of the three checks. Report your results as a rate, not a yes or no. For example, "named in 14 of 40 queries (35%)" is a usable metric. "We showed up once" is not.

Step 4: Calculate the metrics that matter

Turn your log into a handful of numbers you can track month over month:

• Overview coverage: the percentage of your tracked queries that trigger an AI Overview. This tells you how much of your market AI already touches.
• Mention rate: the percentage of Overview-triggering queries that name your business.
• Citation rate: the percentage of those queries that link to your website.
• Accuracy rate: the percentage of mentions that describe your business correctly.
• Share of voice: your mentions compared to the total business mentions across all checked queries.

Example: you track 40 queries. 28 trigger an AI Overview (70% coverage). Your business is named in 7 of those 28 (25% mention rate) and linked in 5 (about 18% citation rate). That tells you AI search is a major channel in your market and that you are absent from three quarters of the answers. That is a specific, fixable gap.

Manual tracking vs automated tracking

Both approaches work. The right choice depends on how many queries you track and how often.

Most businesses start manually for one or two cycles to understand the landscape, then move to automation once the spreadsheet becomes a chore.

Where Found in AI fits

Found in AI is built for exactly this tracking problem. It monitors how AI tools answer the questions your customers ask, shows whether your business is mentioned, and highlights where you are missing. Instead of repeating checks by hand and reconciling screenshots, you see your visibility across the questions that matter in one place and can watch it change over time.

Found in AI works as an AI analyst for marketing: it surfaces the gaps so you know which questions to act on first. It does not replace the work of improving your content and online presence. It tells you where that work should go. If you prefer to start with the manual method above, the metrics and question groups in this guide apply the same way to any tracking approach.

Step 5: Read the results by question group

Raw percentages matter less than patterns. Break your results down by query group and look for these signals.

You appear on brand questions but not on problem questions

This is the most common pattern. Customers who already know you find you, but customers still researching never see you. Fix it with educational content that answers the problem directly, written on pages Google can crawl and quote.

You appear on problem questions but not on local questions

Your content is useful, but Google lacks the local signals to connect it to your service area. Check that your name, address, phone number, service areas, and hours are consistent everywhere they appear, and that your pages state the cities you serve in plain text.

Competitors are named and you are not

Open the sources cited in those Overviews. Note what type of pages are cited: directories, review sites, local news, business websites, forums. The citation pattern shows where AI is finding its information, and therefore where your information needs to exist.

You are named, but the details are wrong

Wrong hours, discontinued services, or the wrong location mean AI is drawing from outdated sources. Update the source, not just your website. Correct business listings, profiles, and any third-party pages carrying stale information.

Step 6: Trace what AI Overviews cite

Each AI Overview shows source links. Treat those links as a map of what Google trusts for that question. For every query where you are absent, record the domains cited and sort them:

• Business websites: compare their page to yours. Do they answer the question in the first paragraph? Do they include specifics such as price ranges, timelines, and steps?
• Directories and review platforms: confirm your profile exists, is complete, and carries recent reviews.
• Local or industry publications: these indicate topics where earning a mention or contributing expertise builds visibility.
• Forums and community sites: these indicate questions where real customer language is shaping the answer.

This step converts tracking from a scoreboard into a to-do list.

Step 7: Close the gaps you find

Tracking only pays off when it drives changes. These actions move the numbers most consistently for local businesses.

• Answer the question first. Put a direct, two-sentence answer at the top of the page, then add detail. AI systems quote content that states the answer plainly.
• Use specifics. Include numbers, timeframes, service areas, and named procedures or systems. "Most water heater replacements take 2–4 hours" is quotable. "It depends" is not.
• Create one page per service and city. A plumber serving three cities needs distinct, useful content for each, not one generic page.
• Add question-and-answer sections. A short FAQ at the bottom of service pages mirrors how customers ask and how AI extracts answers.
• Keep business information consistent. Names, addresses, phone numbers, hours, and service descriptions should match across your website and every listing.
• Build review volume and recency. Reviews are a major input for local recommendations. Ask for them after every completed job.
• Make contacting you simple. A contact form and online booking on your website turn a mention into a lead.

After making changes, wait two to four weeks before judging the effect. AI systems recrawl and refresh at different speeds, so results do not move overnight. Keep checking on the same schedule and compare against your baseline.

Examples of what tracking reveals

A dental practice tracks 35 questions. It appears in 9 of 22 Overviews, mostly for brand queries. For "how long does a dental implant take," a competitor's page is cited. The practice rewrites its implant page with a direct answer, a timeline breakdown, and an FAQ. Over the next two months, it tracks whether the page begins to appear as a cited source.

An HVAC company finds that Overviews for emergency queries name only directory listings. It completes its profiles on those directories, adds recent reviews, and publishes an emergency service page listing the cities it covers and its response process.

A law firm discovers that an Overview about a common legal question describes the firm's practice area incorrectly, drawing from an outdated profile. Correcting that profile fixes the source of the error.

In each case, the tracking data identifies a specific problem and a specific page or profile to fix.

Common tracking mistakes

• Checking once and drawing conclusions. Variation is normal. Test repeatedly.
• Searching while signed in from your office. Your own history skews results, and you probably search for yourself more than customers do.
• Tracking only brand queries. Brand searches are the easiest to win and the least valuable for finding new customers.
• Changing the question list every cycle. You lose the ability to measure progress.
• Ignoring accuracy. Being mentioned with wrong information can cost you customers.
• Stopping after one report. Visibility shifts as AI systems update and competitors publish.

A monthly tracking routine

1. Semana 1: run your full question list and log results.
2. Week 2: review cited sources for queries where you are absent and pick the three biggest gaps.
3. Week 3: publish or update the pages and profiles tied to those gaps.
4. Week 4: retest the affected queries and compare mention rate and citation rate to the previous month.

This cycle keeps the work small enough to sustain and focused enough to produce measurable change.

Google AI Overviews Tracking FAQ

How often should I check whether AI Overviews mention my business?

Check monthly at minimum and weekly if your market is competitive or you are actively changing content. Use the same question list each time so results are comparable, and test each query on at least three separate days to account for variation.

Can I track AI Overview mentions with Google Search Console?

Search Console reports clicks and impressions for pages appearing in Google Search, but it does not give a clean, separate report of whether your brand is named inside AI Overviews. For mention and citation tracking, log the Overviews yourself or use a tool built for AI visibility, such as Found in AI.

Why does my business appear in an AI Overview one day and not the next?

AI Overviews are generated fresh and can change with location, wording, device, and updates to Google's systems. This is why you measure a mention rate across many queries and repeated checks instead of relying on a single result.

How many questions should I track?

Track 20–50 questions that reflect real customer searches across problem, cost, comparison, local, and brand categories. Fewer than 20 produces unreliable percentages, and more than 50 is difficult to maintain manually.

What should I do if an AI Overview shows wrong information about my business?

Find the source the Overview cites, then correct the information there: your website, business listings, and any third-party profiles. Keep your name, address, phone number, hours, and service descriptions identical everywhere, then recheck the query after a few weeks.