Google now answers 83% of branded searches with an AI Overview. Who writes your brand's answer?

For years the branded search was the one part of Google a brand could mostly stop worrying about. Someone typed your name, your site came up first, maybe with sitelinks, maybe with a knowledge panel built from your own facts. Whatever was happening to generic searches, people looking for you by name found you.

A study Ahrefs published on October 2 suggests that's no longer how it works. Ryan Law and Xibeijia Guan analyzed 232.8 million US desktop search results collected between July and September 2026, split into 92.5 million crawls of branded keywords and 140.2 million of non-branded ones. As of September 29, an AI Overview appeared in 82.91% of the branded results.

We've already written about Semrush finding AI Overviews on 88% of commercial searches, the "best X for Y" and "X vs Y" questions shoppers ask before they've picked a brand. The Ahrefs data covers the other end: the moment after someone has already heard of you.

The gap between branded and generic closed in three months

Until recently, Google was noticeably more reluctant to put an AI summary on a branded search than on a generic one. Ahrefs' own research from November 2025 found Overviews far more common on unbranded queries. That gap has nearly closed.

Branded keywordsNon-branded keywords

Show the data

Month	Branded keywords	Non-branded keywords
July 2026	61.3%	73.5%
August 2026	65.1%	74.7%
September 2026	73.3%	78.3%

Source: Ahrefs, AI Overviews Currently Appear for 83% of Branded Searches, October 2, 2026

Branded searches rose 12 percentage points in three months, from 61.3% to 73.3% as a monthly average. Non-branded rose 4.8 points. The difference between the two went from 12.2 points in July to 5 points in September, and the end-of-month snapshot (82.91%) is higher than the September average, which tells you the line was still climbing when the study closed.

A note on scope: these are desktop results in the US, and "branded keyword" means any search containing a brand, so it includes things like a brand plus "login" or "near me." It isn't a CPG-specific sample. We'd expect the direction to hold for food, beauty and household brands, but we can't tell you the exact rate for your category from this study.

For bare brand names, the summary usually sits first

Ahrefs also ran a tighter test. It took 100 brands from Interbrand's Best Global Brands 2025 list and checked the results for the bare brand name at 14 weekly checkpoints. This is the cleanest version of the question: someone types just your name.

At the July 1 checkpoint, 5 of 98 usable results showed an AI Overview. By September 30, 63 of 100 did, with most of the jump coming at the start of September.

Show the data

Checkpoint	Brand names with an AI Overview
July 1 (98 usable)	5
September 30 (100)	63

Source: Ahrefs, AI Overviews Currently Appear for 83% of Branded Searches, October 2, 2026

Position matters as much as presence. Of those 63 summaries, 48 (76.2%) appeared in position one, with an average position of 1.49. In plain terms, on most of these brand searches the first thing on the page wasn't the brand's own site. It was Google's paragraph about the brand.

Ahrefs points back to its own earlier work, which found the top-ranking page gets a 58% lower click-through rate when an AI Overview is present. We'd be cautious about applying that number directly to branded searches. People searching your name often want your site, a store locator or a login, and many will scroll past a summary to get there. But some won't, and the summary sets the frame for everyone who reads it.

The sources Google quotes about you aren't mostly yours

This is the part we'd put in front of a brand team. Ahrefs looked at which outside domains (anything other than the brand's own site) the 100-brand summaries cited.

Show the data

Domain	Share of AI Overviews citing it
Wikipedia	71.9%
YouTube	38.6%
LinkedIn	22.8%
Yahoo Finance	15.8%
Facebook	10.5%
Walmart	7%

Source: Ahrefs, panel of 100 Interbrand brands, July to September 2026. Counts were 41, 22, 13, 9, 6 and 4 AI Overviews respectively.

Wikipedia, cited in 41 of 57, is fairly neutral and at least has a process for corrections. YouTube and LinkedIn are harder. Ahrefs quotes Ann Smarty's concern that the videos cited often aren't from the brand's official account and "may very well include outdated product information." Chris Long, whose test of 100 enterprise brands prompted the study, put it more bluntly on LinkedIn: outside sources "can now control the narrative of your brand directly in the search results."

The one retailer on the list caught our eye. Walmart was cited in 4 of the 57 summaries. These were global brands, not grocery brands specifically, but for a CPG company it's a reminder that your retailer product pages are part of how Google describes you, not just where people buy you.

There was some reassurance on competitors. Of 58 summaries where Ahrefs could see the full text, only 2 discussed rivals: Pepsi's named Coca-Cola in describing the cola wars, and Intel's named TSMC and NVIDIA. Only one competitor-owned site appeared across 57 source lists. Google doesn't seem to be sending people who searched for you to someone else, at least not in this panel.

Why this lands differently for a CPG brand

The brands in Ahrefs' panel are huge, with Wikipedia pages maintained by many editors and decades of coverage. Most CPG founders we talk to have none of that. That cuts two ways.

If Google has less to draw on, a smaller brand may simply not get a summary for its name yet. That's the optimistic case. The less comfortable case is that when it does, the summary is assembled from whatever thin material exists: a launch video someone else uploaded, a three-year-old LinkedIn description that still lists your original flavor lineup, a retailer page with the old pack size. Google's documentation on AI features says there are no additional requirements to appear in AI Overviews beyond normal Search, and the controls it lists are the usual snippet controls like nosnippet and max-snippet. Nothing there lets you edit what an Overview says about you. The material it draws from is the only lever.

The other difference is what people are searching for. A shopper typing a beverage brand's name may want to know where to buy it, whether it has caffeine, or whether it's the brand they saw on a friend's counter. Those are fact questions, and a summary built from stale facts gets them wrong in a way that costs a sale.

What we'd do with this

None of this needs new tools. It needs someone to look.

1. Search your bare brand name, signed out, and screenshot the result. Note whether there's an AI Overview, where it sits, and every claim it makes. Our new tip on checking your branded AI Overview walks through it.
2. Open every source it cites. Fix the ones you control first: your official YouTube channel's video descriptions and titles, your LinkedIn company page (what you make, where you're sold, founding year), and your About page.
3. Look at the videos that aren't yours. You can't edit a reviewer's video, but you can publish an official one that answers the same question with current facts, and describe it in plain words so Google can match it.
4. Treat Wikipedia and Wikidata carefully. If an entry exists and is wrong, suggest a sourced correction on the talk page rather than editing it yourself. If none exists, tip 19 covers whether you're likely to qualify.
5. Check your retailer pages. If Google can cite Walmart about a global brand, it can cite a Walmart or Target listing about yours. Make sure the pack sizes, claims and ingredient lists match your site, which is the one-entity problem in practice.

Then repeat monthly. Summaries change, and the Ahrefs data shows they can change fast. We keep a simple log of the same handful of brand and product questions each month, across Google and the main assistants. Some teams pay for that; XENA Intelligence is one company that runs those shopper-style questions through ChatGPT, Gemini, Perplexity and Amazon's assistant and scores how each product shows up, and its SearchPanel shows the answers themselves. A shared doc with dated screenshots also works.

What the study doesn't tell you

It doesn't tell you how many people read the summary and stop there, or how that differs for a brand-name search versus a generic one. It's desktop-only, so mobile, where most grocery and beauty browsing happens, may look different. And it doesn't break results down by industry, so we can't say whether a snack brand is more or less likely to get a summary than a software company.

What it does show clearly is that the branded search is no longer a page you own by default. The summary at the top is written from what the rest of the web says about you, and the biggest contributors are sites most brand teams haven't looked at in a year.