Google's AI Overview now sits on 18 of 21 big packaged food brand names. It had 2 in August

Type your own brand name into Google this weekend. There is a good chance the first thing on the page is no longer your website. It is a paragraph Google wrote about you.

On November 6, 2026 BrightEdge published ten weeks of tracking on exactly that. It followed 282 bare brand and product names, things like a company or product name with nothing added, on Google US desktop, every week from week 34 to week 43 of 2026 (mid August to October 24). The share of those names showing a Google AI Overview went from 19% to 64%. BrightEdge calls it a 3.4-fold rise, and notes that over the same weeks the count of keywords showing an overview across its whole tracked list grew 47%. Brand names moved much faster than everything else.

We covered a different measurement of the same trend in October, when an Ahrefs study put AI Overviews on 83% of branded searches. BrightEdge's number is lower, and it explains why: its list is 282 bare names chosen by analysts, with 67 restaurant brands and 59 names that return a local map pack, the two groups with the least coverage. Leave out restaurants and the closing figure is 78%. Different lists, different numbers, same direction.

The climb, week by week

The rise was not a single switch. BrightEdge published selected weeks.

Show the data

Week	Share with an AI Overview
Wk 34 (mid Aug)	19%
Wk 36	31%
Wk 38	50%
Wk 40	57%
Wk 42	74%
Wk 43 (Oct 24)	64%

Source: BrightEdge, Nov 6, 2026. Counts were 53, 88, 140, 162, 208 and 180 of 282 names. Presence only: not citations, content or clicks.

Note the last two bars. The count hit 208 names in week 42 and fell back to 180 a week later, and BrightEdge says individual names dropped out and came back. It warns against reading 64% as a settled level, and points out that it saw the same spike and pull-back pattern in ecommerce queries in 2025. If you check your brand once and see no overview, that is a reading, not a verdict.

Packaged food was one of the fastest movers

For CPG, the most useful table in the report breaks out the 130 tracked brand names with more than 100,000 monthly searches by category.

Week 34Week 43

Show the data

Category	Week 34	Week 43
Devices and hardware (20)	5	20
Packaged food and beverage (21)	2	18
Games and entertainment (17)	2	15
Finance and payments (23)	1	13
Software and platforms (21)	1	11
Restaurants (24)	1	2
Retail and apparel (4)	0	0

Source: BrightEdge, Nov 6, 2026. Number of names in each category in brackets. Google US desktop.

Packaged food and beverage went from 2 of 21 names to 18 of 21. That is the second-biggest move in the table after devices, which reached all 20. If you run a food or drink brand with meaningful search demand, the base case is now that Google answers a search for your name with its own summary. Retail and apparel shows zero, but there are only four names in that row, so we would not read anything into it.

BrightEdge does not name the packaged food brands, and it does not say what the overviews said. It measures presence only. That limit is important, and we come back to it below.

The biggest brands arrived last

A natural assumption is that famous brands are safe, because Google has plenty of its own information about them. The data says the opposite happened in order: smaller names got overviews first, and large names caught up.

Monthly searches for the name	Names	Week 34	Week 43
Over 1 million	32	0 of 32 (0%)	18 of 32 (56%)
250,000 to 1 million	41	3 of 41 (7%)	23 of 41 (56%)
100,000 to 250,000	57	9 of 57 (16%)	38 of 57 (67%)
Under 100,000	152	41 of 152 (27%)	101 of 152 (66%)

Source: BrightEdge, November 6, 2026, Google US desktop.

Weighted by search volume, BrightEdge estimates the 50 highest-volume names carrying an overview went from about 5.9 million monthly searches to about 64.8 million, roughly 11 times more. It flags that one term, Google Classroom, accounts for 17.5 million of that, so treat it as an order of magnitude. BrightEdge also says a rollout tiered by search volume would fit this shape, but so would other explanations, and it found no public announcement of a branded-query change. We agree with that caution.

The version of the name matters

One small example in the report deserves attention from anyone who owns a brand with a nickname. "American Express" carries an overview and "amex" does not. BrightEdge's point: checking one version of a name and assuming the rest is not an audit.

CPG brands are full of these. The full brand name and the short one. The product line said with and without the parent brand. The flavor or scent name people actually type. The misspelling. Each can trigger a different result, a different summary, or none at all.

Intent matters too. BrightEdge found terms containing "stock" now show an overview 84% of the time and terms containing "near me" 13%. Of the 59 tracked names that also show a local 3-pack, 20% had an overview, against 75% of the other 223. Restaurants were the most protected group, and 40 of those 59 local names are restaurants. For a packaged goods brand sold through retailers, that local shield mostly does not apply.

Why this connects to who gets cited

An overview on your brand name is a summary assembled from sources. Your website is one possible source among many. A separate study published this week by Babel42 found that across 2,053 brand recommendations in Google AI Overviews, the overview cited the brand's own site 26.2% of the time, and that consumer categories were cited on their own site far less than software or finance. We go through that study in today's data post.

Put the two together and the shape of the problem is clear. Google is now describing most brands on their own name search, and when it describes consumer brands, it often uses other people's pages as the evidence. A brand that has never read its own overview does not know which review, retailer page, video or forum thread is currently defining it.

It is also worth noting that Google has been changing where these brand overviews appear. We wrote earlier this month about brand overviews moving into the right-hand column. And on commercial searches, AccuRanker measured overviews jumping from 26% to 51% during September, which we covered in the commercial searches post. The brand search is just the newest place the same change has reached.

A brand-name audit for a CPG team, in one afternoon

BrightEdge closes its report with an audit. Here is our version for a food, drink, beauty or household brand. None of it needs software.

1. List every version of the name. Full brand name, short name, parent brand plus product line, your top three product names, the flavor or scent names people search, and common misspellings. Ten to twenty terms is normal.
2. Search each term in a clean browser, signed out, in the US. Record whether an overview appears. Do it again next week. BrightEdge saw one week move by 28 terms, so a single check misleads.
3. Read the overview the way a shopper would. What does it say you make, who it is for, what it costs, where it is sold, what it contains? Flag anything wrong, out of date or missing, especially formula, allergen, size and price claims.
4. List the cited sources and sort them. Yours, retailer, review site, news, Reddit, YouTube, competitor. This is the list of pages currently shaping your description.
5. Fix facts at the source, not in the overview. Correct your own pages first, then ask the retailer, reviewer or publisher carrying the stale fact to update it. The tip on fixing a wrong answer has the steps.
6. Keep the baseline. Save screenshots and the source list with dates. If the overview changes after you fix a source, that is the evidence your fix worked.

When the overview gets the basics right, the next question is whether assistants other than Google say the same thing. A brand that looks fine in AI Overviews can be described differently in ChatGPT or Perplexity, which read different sources. That cross-assistant check is what XENA Intelligence focuses on for consumer products, with SearchPanel reporting per product; the manual version is the same 20 terms asked as questions in each assistant once a month.

What the data does not show

BrightEdge is clear about its limits, and so should we be. It measures whether an overview appears, on desktop, in the US. It does not measure whether your site is cited inside it, what it says, or whether it changes clicks to your site. The list of 282 names was chosen by analysts and includes many restaurants, which is why it reads lower than other studies. And BrightEdge sells software for this audit, which the report mentions.

What the data does show is enough to act on. In ten weeks, the AI Overview went from a minority of brand searches to most of them, and packaged food brands were among the fastest to get one. If your team has a brand search report, it should now include the overview text and its sources, not only your position and clicks. See tip 36 for the five-minute version of this check.