Google now shows AI Overviews when people search your brand name. Here's what to check.

When someone types your brand name into Google, they already know who you are. They might be about to buy, renew, sign a contract or check whether you're trustworthy. Usually, the thing they're looking for is your own website.

That is changing. Google now shows an AI Overview (the AI-written summary at the top of some search results) on a large share of searches for brand names. So the first thing many people read about you is no longer your homepage. It is Google's summary of what the web says about you.

This piece covers what the data shows, why a branded AI Overview needs more care than a normal one, and a step-by-step audit you can run on your own brand in an afternoon. It ends with a scoring worksheet and a list of fixes, sorted by where the problem actually lives.

What the data shows about AI Overviews on brand searches

The clearest public numbers come from Ahrefs, an SEO software company. In a study published on October 2, 2026, Ahrefs analyzed 232.8 million US desktop Google search results crawled between July and September 2026. Of those, 92.5 million were for branded keywords (searches that include a brand's name).¹

Here is what they found:

Ahrefs also tracked a smaller, separate group: 100 brands from Interbrand's Best Global Brands 2025 list. They searched each brand name at 14 weekly checkpoints between July and September. On July 1, only five of 98 usable results had an AI Overview. By September 30, 63 of the 100 brand-name searches had one.¹

A note on the data: these two sets of numbers measure different things, so they don't match. The big crawl covers all kinds of branded keywords, many of them longer searches like "brand + login" or "brand + pricing." The 100-brand panel looks only at the bare names of very large global brands. Both point in the same direction, though. AI Overviews on brand searches went from rare to common within one quarter. Also, all of this is US desktop data. Mobile and other countries may look different, which is one reason to check your own brand rather than assume.

Which sources Google leans on to describe a brand

An AI Overview shows links to the pages it drew on. Ahrefs looked at those links in the big-brand AI Overviews where the source list was visible. The results show that the most-cited sources were mostly not the brands' own websites.¹

Ahrefs described Wikipedia as "a fairly neutral source of information for most brands."¹ That's fair. But it also means a page you don't control often sets the first impression.

Ahrefs' own summary of the risk was blunt:

"competitors/external sources can now control the narrative of your brand directly in the search results."

Ahrefs, October 2, 2026.¹

Only two of 58 answers in the panel (3.4%) discussed competitors. So for big brands, the main issue is not rivals sneaking into your branded results. The issue is that third-party pages, some of them old, are deciding what Google says about you.

Why a branded AI Overview matters more than a normal one

A normal AI Overview answers a general question, like "how do I fix a leaky tap." A branded one answers a question about you. That makes it different in three ways.

The reader is close to a decision

People who type your exact name have usually heard of you already. They're often checking prices, looking for a login page, reading reviews or deciding between you and someone else. So the summary they see lands at a moment that matters for sales and trust.

Old or negative facts can get pulled to the top

Google says plainly that "AI Overviews can and will make mistakes."² The summary is built from whatever pages Google finds useful. If those pages carry an old price, a former CEO, a product you retired or a complaint thread from three years ago, that can end up in the summary. And because the summary sits at the top, it can look more official than the source it came from.

It may take some clicks away from your own site

Be careful with this one, because the evidence is not branded-specific yet.

• Ahrefs compared 300,000 keywords and found that position-one click-through rate (the share of searchers who click the top result) was about 58% lower when an AI Overview was present. That study used informational keywords, not brand names.³
• Pew Research Center studied the browsing data of 900 US adults in March 2025, covering 68,879 Google searches. When an AI summary appeared, people clicked a traditional search result in 8% of visits. Without one, they clicked in 15% of visits. Clicks on links inside the summary itself happened in just 1% of visits.⁴

Ahrefs' branded study says these overviews are potentially reducing clicks to brands' own sites, not that they definitely are.¹ That's the right level of caution. Someone looking for your login page will probably still click through. Someone checking whether you're legit may read the summary and stop there. We haven't found an independent click study focused only on branded searches yet. So watch your own numbers instead of assuming a fixed loss.

Google, for its part, says that clicks from results pages with AI Overviews "are higher quality (meaning, users are more likely to spend more time on the site)."⁵ So fewer clicks doesn't automatically mean fewer good visits. You need your own data to know which is true for you.

How Google builds these answers, and what you can and can't control

Google's own documentation explains a few things that shape the audit below.

• It searches more than once. Google says AI Overviews and AI Mode may use a "query fan-out" technique. This means Google runs several related searches across subtopics and data sources to build one answer.⁵ In other words, a search for your name may quietly pull in pages about your pricing, reviews and news too.
• There's no special entry ticket. To be shown as a supporting link, "a page must be indexed and eligible to be shown in Google Search with a snippet". Google adds: "There are no additional technical requirements."⁵
• They don't always appear. Google says AI Overviews "are only shown when our systems determine that it is additive to classic Search, and as such, often don't trigger."⁵
• You can limit what's taken from your own pages. Google lists the nosnippet, data-nosnippet, max-snippet and noindex controls for this.⁵ These are instructions you add to a page's code that tell Google not to show some or all of its text.

Search Console (Google's free tool for site owners) also has a "Search generative AI control." Since August 31, 2026, it has been available to all websites worldwide. It's a setting that removes your site from AI Overviews, AI Mode and generative AI features in Google Discover.⁶ But it does not remove the AI Overview from your brand searches. Google's help page is clear about this:

"Content from other sites will still be available in those features, and it may appear similar to yours."

Google Search Console Help.⁶

So opting out, or hiding text with nosnippet, only removes your version of the facts. The summary still appears, built from Wikipedia, reviews and other sites. For a branded search, that's usually the opposite of what you want. The better fix is to make sure the accurate version of your facts is easy to find on your own site and on the third-party pages Google already trusts.

The branded AI Overview audit, step by step

This is the audit we suggest for brands and agencies. It takes a few hours the first time and less after that. You need a spreadsheet, a browser and the worksheet in the next section.

Step 1: Build your list of branded searches

Don't only check your bare brand name. People attach words to it, and each version can show a different summary from different sources. Start with these groups:

Aim for 15 to 30 searches in total. Add searches your sales or support team hear often. If your brand name is also a common word, add a category word ("Brand software," "Brand shoes") so you're testing the right thing.

Step 2: Search in a clean, repeatable way

AI Overviews can change between people, places and devices. So keep your setup the same every time you run the audit.

• Use a private or incognito window, signed out of Google.
• Note the country and city you're searching from. If you sell in several countries, repeat key searches for each one.
• Check both desktop and mobile. The Ahrefs data above is desktop only, so don't assume mobile matches.
• Write down the date and time. Take a full-page screenshot of each result.
• If an AI Overview has a "Show more" button, open it before recording, so you see the full text and source list.

Step 3: Record the same details for every search

For each search, fill in one row of your spreadsheet:

1. Did an AI Overview show? Yes or no. Also note if it showed only after you clicked something.
2. Where on the page? Above everything, below an ad, or lower down. Ahrefs found most big-brand overviews sat in position one.¹
3. Which sources did it cite? List every linked site. Mark each one as owned (your site), controlled (your official LinkedIn, YouTube or review-site profile), or third party.
4. Are the facts right? Check every claim: founding year, location, leaders, products, prices, customer numbers, awards.
5. Are the facts current? A fact can be true but old, like a price from two years ago. Mark these separately from flat-out errors.
6. What's the tone? Positive, neutral, mixed or negative. Note the exact words that set the tone.
7. Where did each wrong or old fact come from? Click the cited links and find the sentence. This is the most useful column, because it tells you what to fix.

Step 4: Score each result

Use the worksheet below. Scoring makes it easy to compare this month with next month, and to see which query groups need work first.

A simple scoring worksheet

Give every search that shows an AI Overview a score out of 8, using four checks worth up to 2 points each.

Then add it up:

• Score per search = Accuracy + Freshness + Sources + Tone (out of 8).
• Score per query group = the average of the searches in that group.
• Overall score = total points ÷ (number of searches with an AI Overview × 8).

Keep a separate count of how many of your searches showed an AI Overview at all. That number tells you how big the issue is. The score tells you how well it's going.

A filled-in row of your sheet might look like this:

An illustrative example: a mid-size software brand

This is an illustrative composite, not a real client or study. "Fernbrook" and "Quillstack" are invented names. The example is built to show how the audit works, using the kinds of problems the audit is designed to catch.

Fernbrook is a payroll software company. Its team runs the audit on 20 branded searches, signed out, on desktop, from one US city. Twelve of the 20 show an AI Overview. Here are five of them:

Across all 12 overviews, Fernbrook scores 58 out of a possible 96. That's 60%. But the average hides the real story. The bare brand search and reviews are in good shape. Pricing, comparison and trust searches are where it's losing people, and those are exactly the searches buyers make right before they decide.

The "where it came from" column gives the team its to-do list:

1. The pricing page shows prices only after a sign-up form. So Google can't read them, and quotes an old review instead. Fix: put current starting prices in plain text on the pricing page, with a "last updated" date.
2. The mobile app has no page of its own. It's mentioned once in a press release. Fix: publish a clear product page, and email the comparison blog's author with the launch date.
3. The outage has never been explained on Fernbrook's own site. So the forum thread is the only source. Fix: publish a short, honest status-history or incident page that says what happened and what changed.
4. The Wikipedia article still lists the old CEO. Fix: post an edit request on the article's talk page with a link to the press release, following Wikipedia's conflict-of-interest rules (more on this below).

What to fix, sorted by where the problem lives

You can't edit an AI Overview directly. You can only change the pages it learns from, and tell Google when something's wrong. Here's where to look, depending on what your audit found.

A word on Wikipedia

Since Wikipedia shows up in so many big-brand overviews, it's tempting to go and fix your page. Be careful. Wikipedia's conflict-of-interest guideline says: "Paid advocates are very strongly discouraged from direct article editing, and should instead propose changes on the talk page of the article in question."⁷ It also says: "You must disclose your employer, client, and affiliation with respect to any contribution for which you receive, or expect to receive, compensation."⁷

So the right move is an edit request on the talk page, with links to reliable sources (news coverage, official filings, your own press release for simple facts like a CEO change). The same care applies to Wikidata, the structured database linked to Wikipedia. And if your brand doesn't have a Wikipedia article, don't create one just to influence AI answers. Focus on the sources you control.

Knowledge Panels and Google feedback

A Knowledge Panel is the information box Google shows beside some brand searches. Google says these panels draw on many sources across the web, and update automatically as that information changes. Google also considers changes suggested by the brand itself and by general user feedback.⁸ Anyone can use the feedback link on a panel to suggest an edit.⁸ If you represent the brand, you can also claim the panel. You do this by clicking "Claim this knowledge panel" and signing in to one of the official sites or profiles Google lists. Once verified, you can suggest changes directly.⁹

For the AI Overview itself, Google's help page says to click the thumbs-down icon if a summary is "unhelpful, inaccurate, biased, or otherwise problematic," then click "Report a problem."² Treat this as a backup, not the main fix. If the wrong fact is still on a trusted page, it can come back.

Three mistakes to avoid

1. Opting your whole site out to "make it go away." As covered above, the overview will still appear. It will just be built from other people's pages, not yours.⁶
2. Hiding your best facts with nosnippet. This stops Google from quoting your own clear wording. Save it for text you really don't want shown, like legal fine print.
3. Only checking your bare brand name. In our composite example, the bare name was fine. The problems were in pricing, comparison and trust searches. Those are easy to miss if you only search your name.

How to track this over time

Run the same list of searches once a month, in the same conditions, and keep the scores in one sheet. AI Overviews can change week to week, as Ahrefs' weekly panel showed. So one snapshot isn't enough.

Search Console can help a little. Google says traffic from AI features like AI Overviews and AI Mode is counted within the "Web" search type in the Performance report.⁵ There's also a newer generative AI performance report. It shows how often links to your site appeared in AI Overviews and AI Mode.¹⁰ But it only reports impressions (how many times your link was shown). You can group the data by page, country, device and date.¹⁰ It doesn't show which searches triggered the overview, and it doesn't show clicks. That's why the hands-on audit still matters.

A practical pairing:

• Use Search Console to watch whether your About, pricing and product pages start getting AI impressions after you fix them.
• Watch branded clicks to your homepage in the normal Performance report. If they fall while branded impressions stay flat, read the AI Overview for those searches.
• Use the monthly audit to check what the summary actually says.

If you want to go beyond Google, the same idea applies to ChatGPT, Gemini and Perplexity. We cover that in our guide to AI brand visibility, and the tone side in our piece on AI brand sentiment.

See what AI says when people search your name

A branded AI Overview is a summary of your reputation, written by Google from other people's pages. Ripplix checks how AI search tools describe your brand, which sources they rely on, and where the facts are wrong or out of date, so you know exactly which pages to fix first.

Get your free AI Visibility Report →