AI Overviews: 83% of Brand Queries, the Ahrefs Study

83%: What Does the Ahrefs Study Really Measure?

82.91%. That is the share of US Google desktop results, for brand searches, that showed an AI Overview on September 29, 2026. The figure comes from a study published by Ahrefs on October 2, 2026 (Ryan Law, with Xibeijia Guan).

232.8 million crawls analyzed from July to September. 92.5 million on brand keywords. 140.2 million on non-brand keywords. This is not a test on ten queries on a Friday night.

Why does this figure deserve your attention? Because the brand query was the last safe haven. The ground where your site ranked first without effort, with a comfortable click-through rate. Someone types your name, they want your site. Full stop.

That full stop is fading. Google now writes the answer, then cites sources. Sometimes yours. Often someone else’s.

It all started with a homemade test by Chris Long: an AI Overview on 93 of the 100 company brand names tested. Ahrefs wanted to verify at scale. Scale changes the number (83% versus 93 out of 100), not the direction.

A methodological note before moving on. « Brand keywords » here means searches that contain a brand, not just the bare name. The panel of 100 brands we will see below is a separate sample, limited to bare names, like « bluehost ». Two datasets, two readings.

A field observation, not a statistic: I type my clients’ names into Google every week. It’s a diagnostic reflex. Since the start of the autumn, a generated answer appears more and more often before their site. Ahrefs puts a number on what my screens were already showing.

One last point, in the interest of honesty. The study covers the United States, in English. Nothing says France shows the same rate today. Still, a trend is easier to prepare for before it arrives. Better to measure on your own site first.

Entre juillet et septembre, les AI Overviews se sont massifiées sur les recherches de marque. L’écart avec les requêtes génériques se réduit de 12 points à 5 seulement : les marques ne sont plus un sanctuaire.

What Changed Between July and September 2026?

One quarter. That is all it took.

Ahrefs publishes the monthly averages of keywords that trigger an AI Overview. Read the left column.

On brand: +12 points between July and September. Ahrefs calculates roughly 20% relative growth compared with the July rate. On non-brand: just +4.8 points.

The gap between the two query families has narrowed. 12.2 points in July. 5 points in September. In its previous study on AI Overview triggers, in November 2025, Ahrefs found these blocks were far more frequent on unbranded queries. That hierarchy no longer holds. The samples differ, so the comparison is an order of magnitude, but the direction is clear.

Wondering why you read 73.3% in the table and 82.91% in the title? The first is a monthly average. The second is the snapshot of September 29. When the curve climbs fast, the latest picture mechanically exceeds the monthly average. In fact, in the panel of 100 major brands, Ahrefs notes that AI Overview presence exploded from the start of September.

This is a tipping point, not a gradual drift. In a few weeks, the brand keyword went from protected zone to testing ground.

My hypothesis, to be read as such: Google opened up brand queries once its generated answers became reliable enough to answer « who is this company? ». Ahrefs makes no such claim. The numbers, though, are there.

What should you do with this information? Stay calm. You gain by looking at your own quarter: your July brand clicks against your September ones. If the slope looks like Ahrefs’, you know where you stand. If it differs, you hold data nobody else has. Either way, you move forward with concrete facts.

And what if your brand was already among the 83% and you hadn’t seen it?

Why Position 1 Changes Everything for Your Brand Name

5 out of 98. 63 out of 100.

These two numbers tell the story. Ahrefs tracked the bare names of 100 brands from the Interbrand Best Global Brands 2025 ranking, at 14 weekly checkpoints between July and September. On July 1: 5 AI Overviews among 98 usable results. On September 30: 63 out of 100.

Even more important. Of those 63 AI Overviews, 48 appeared in position 1. That is 76.2%. Average position: 1.49.

An AI Overview can appear in several places on a results page. But Ahrefs observes a drop in click-through rate when an AI Overview is present, with the strongest impact when the block takes the top spot on a brand query, in place of the brand’s domain.

What matters is what shows first on the screen. Not your ranking.

Three levels of exposure stand out:

• Your site is cited in the AI Overview — you stay in the conversation, you lose part of the click.
• A third-party source dominates — another site tells your brand’s story, in its own way.
• No AI Overview — the historical situation, now rare on the bare names in the panel.

Ahrefs adds a nuance: Google could decide to move these blocks lower on the page. We don’t know whether that will happen. The September data mostly show that they occupy the top of the page.

Counterintuitive, and this is what I take away: ranking number 1 on your own name no longer means appearing first on the screen. For years, the brand query served as proof of SEO health. Today, a perfect ranking can coexist with falling clicks.

So, what should you bet on? On what Google reads to build its answer. Your site, of course. But also everything that talks about you elsewhere. That is the topic of the next section.

Qui résume votre marque? Wikipedia (71,9%), YouTube (38,6%) et LinkedIn (22,8%) dominent les citations des AI Overviews. Chaque overview pouvant citer plusieurs sources, le total dépasse 100%.

Wikipedia, YouTube, LinkedIn: Who Speaks for Your Brand?

The real threat of an AI Overview on a brand query is not the block itself. It is the source it cites.

Ahrefs examined the external citations (outside the brand’s own site) in its panel of 100 brands. Here is the ranking.

The percentages add up to more than 100%: a single AI Overview can cite several domains.

Wikipedia dominates with nearly 72% of external citations. Ahrefs considers it fairly neutral for most brands. The next two raise a different question. YouTube and LinkedIn are platforms where a brand controls its narrative far less than on its own pages, according to the study.

Ahrefs illustrates the case with the query « bluehost », where YouTube is heavily cited in the AI Overview. A video hosted on a third-party platform then becomes the brand’s voice.

Note also the presence of Yahoo Finance, Facebook and Walmart. A retailer talking about a brand it doesn’t own: that matters. Ahrefs also points out that the ultimate risk is steering the user toward unofficial sources or, in the worst case, toward competing brands.

Apply this to a small business. Three examples of how to read it, to verify on your side:

• Is your LinkedIn page empty or outdated? It can become the source of a Google answer.
• Is there no official video carrying your brand name? A third party’s video takes the spot.
• Does your Wikipedia page (if it exists) contain outdated information? It feeds the answer.

What matters is your consistent presence on the three or four surfaces Google reads first to summarize a brand.

In my view, this is the hidden good news of the study. These surfaces are identifiable. Measurable. You gain by treating them as extensions of your site, with the same editorial care.

Which source is talking about you right now, and have you already read it?

What About CTR by Industry? What the Study Says, and What It Doesn’t

Let’s be precise. The Ahrefs article publishes no click-through rate by industry. It establishes an appearance frequency (82.91%, 73.3%, 63 out of 100), a position (1.49 on average) and cited sources. It recalls that its earlier research shows a drop in click-through rate when an AI Overview is present.

So I won’t invent a « -XX% for e-commerce » for you. That figure does not exist in the source. What does exist is your Search Console, and it holds the answer for your site.

This signature has a precise meaning. Your site has held its ground. It appears just as often, at the same rank. What changes is what the user does after reading the block above.

For an industry-level reading, segment your brand queries by intent. My starting hypotheses, to test rather than to believe:

• E-commerce: « brand + product » and « brand + reviews » are likely to be summarized in the block. Direct purchase queries deserve separate tracking.
• SaaS and services: « brand + pricing » and « brand + alternative » are queries where the generated answer may cite a third-party comparison site.
• Local and neighborhood businesses: opening hours, address, reviews. Information Google already knows how to display.

No percentage here, on purpose. On the topic clusters I build, the method is to segment first, interpret second. An industry figure borrowed from another site would have you steering with the map of another territory.

The opportunity lies on the other side of the question. A generated answer satisfies « who is this company? ». It satisfies less « can I buy, book, request a quote, and at what exact price? ». Pages that answer this second family of questions stay valuable.

Keep this sentence in mind: the brand query isn’t disappearing, it’s moving upmarket. The clicks that remain are closer to the decision.

Does your July-to-September data already tell this story?

What Can You Do This Week to Take Back Control of Your Name?

A plan in six moves. Each one needs no media budget.

1. Audit your ten brand queries. Type them in a private browser window. Note: AI Overview present or not, position of the block, cited sources, your site cited or missing. A spreadsheet is enough. It is your starting snapshot.

2. Check what Wikipedia says. First external source cited, with 71.9% of the panel’s external citations. If your brand appears there, make sure the information is accurate and sourced, following community rules. If it doesn’t, don’t force anything: build the notability that makes a page legitimate.

3. Publish a structured official YouTube channel. Second source cited, with 38.6%. Videos that carry your brand name and answer the questions people ask about you: who you are, what you offer, how it works.

4. Polish your LinkedIn page. Third source, with 22.8%. Complete description, recent activity, information consistent with your site.

5. Build your brand pages on your own domain. About, pricing, FAQ, case studies, comparisons. Every question an AI Overview summarizes should find its best answer on your site. I build these architectures — brand semantic clusters — so Google ranks your source first.

6. Set up monthly tracking. Same ten queries, same grid, same day of the month. Add the brand click curve from your Search Console. In three months, you have a trend of your own, not a US average.

That last sentence is a field observation, not a statistic. I give it to you as such.

The Ahrefs study shows it: your brand isn’t secured once and for all. You now know where to look. The surfaces to polish can be counted on one hand.

If someone types your name tonight, who answers them first: your site, or Wikipedia, YouTube and LinkedIn?

Stéphane Jambu

SEO & AI Engineer

I build growth systems / AI / Neuroscience | 650+ clients · 80 LinkedIn testimonials · 30 years of expertise · 15 years of systems running without me.

Follow on LinkedIn