AI & Reputation
November 2026
The AI Search Brief
The Status Labs GEO Desk · Austin

Google told investors that its AI answers are drawing more searches and earning less per search, and Alphabet shares then had their worst day since 2004. OpenAI says ChatGPT now handles about a fifth as many searches as Google, and both companies are selling ads inside their answers. On brand searches, Google has started placing an AI Overview where the knowledge panel used to sit, and Ahrefs data from late September puts an overview on 83% of branded results.

On our radar
  • Google's AI answers bring more searches and less money per search, and Alphabet has its worst day as a public company.
  • OpenAI puts ChatGPT's search volume at about a fifth of Google's.
  • GPT-6 reaches every ChatGPT user, and a week of ChatGPT visits goes missing from analytics.
  • Google tests an official-sites box in AI answers, and Reddit starts closing its doors to logged-out readers.
  • On brand searches, AI Overviews take over the knowledge panel's spot.
  • New research on what gets a page cited by ChatGPT: the search it runs counts more than the page.
ONE BIG THING

Google's AI answers bring in more searches and less money per search

Alphabet reported on Oct. 28 that Google's Search revenue grew 12% in the third quarter, to $63.35 billion, short of the 15% analysts expected and down from 17% growth in the second quarter. On the earnings call, chief business officer Philipp Schindler said "the strong growth in AI Mode queries had an impact because of the lower monetization we were seeing in AI Mode." The next day Alphabet's Class A shares fell 12.6%, their worst day since the company began trading in 2004, erasing about $482 billion in market value.

Schindler also said Google is taking AI Mode ads to international markets for the first time and adding multi-turn ads in the U.S., and that click-through rates on sponsored product listings in AI Mode are already comparable to those of organic listings. Sundar Pichai said AI Mode queries run about three times longer than traditional searches, and that Google's query growth is the highest it has been in five years.

At a dinner in New York, OpenAI's ChatGPT consumer product lead, Adam Fry, said ChatGPT's search volume is "roughly one-fifth the volume of queries of Google," counting any prompt that sends ChatGPT to the web as a search. OpenAI did not say where its Google figure came from, so the comparison is unverified. Fry also said ChatGPT's ads passed $1 billion in annual recurring revenue within 200 days.

On both platforms, a paid recommendation can now sit inside the same response that describes a company, next to the sources the model picked on its own. Any audit of how an engine presents a brand now has to separate what the brand earned in the answer from what a rival paid to put there.

THE PLATFORMS

GPT-6 reaches every ChatGPT user, and a week of ChatGPT traffic goes uncounted

OpenAI brought GPT-6 to every ChatGPT tier on Oct. 7 and 8, pitching it to "more than 1.2 billion people who use ChatGPT each week." A feature called Intelligent UI builds answers from charts, forms, buttons and small tools. OpenAI says GPT-6 Instant starts answering web-search questions 44% sooner than the model it replaced and makes better decisions about when to look something up.

Inline links in ChatGPT answers stopped carrying the utm_source=chatgpt.com tag or a referrer after the rollout, so the visits showed up in analytics as Direct, Search Engine Journal reported on Oct. 15. On about 500 pages where ChatGPT drove at least 5% of visits, Adobe's Cedric Huesler counted roughly 31,000 fewer ChatGPT visits a day from Oct. 9 to 13 than the week before, while Direct visits rose by about 26,000 a day. Huesler said OpenAI confirmed a bug on Oct. 14 and that tagged clicks recovered from Oct. 15.

Any report covering Oct. 8 to 15 needs a note before anyone reads it as lost AI traffic. Huesler recommends counting the UTM tag, since most ChatGPT clicks still arrive without a referrer.

Google adds links to its AI answers and tests a box for official sites

Google said on Oct. 15 that it has added many more inline links to AI Overviews and AI Mode and shows them higher in responses more often. It is testing a dedicated section at the top of some answers for the official pages that complete a task, each with a note on why the link is useful. News carousels that highlight Preferred Sources are reaching more countries and languages; Google says the number of people using Preferred Sources has grown more than sevenfold since the start of the year.

For navigational and task questions, the test would put a company's own site above the summary. Google gave no data on whether the added links send more clicks. Separately, Gemini 4 Argon reached AI Mode for Google AI Pro and Ultra subscribers on Oct. 13.

Reddit starts walling off the threads AI answers lean on

Reddit said on Oct. 29 that it will move to a primarily logged-in experience, with tests outside the U.S. in the fourth quarter. The company said the shift protects its data from AI scrapers and positions it for licensing deals with model makers. Search referrals fell by more than 1 million a week on average in October, and U.S. daily active users slipped to 50.7 million in the third quarter from 53.2 million in the second.

Semrush's June 2026 AI Visibility Index found that ChatGPT frequently relies on community and reference sites such as Reddit and Wikipedia. Once more threads sit behind a login, what an engine can read about a product on Reddit may depend on whether its maker has signed a deal.

GEO & REPUTATION

On brand searches, AI Overviews move into the knowledge panel's spot

Google has moved AI Overviews for many brand-name searches into the right-hand column on desktop, Search Engine Roundtable reported on Nov. 2, with examples including Apple, Zillow and Coke. Editor Barry Schwartz wrote that Google is in effect replacing brand knowledge panels with AI Overviews. Google has not commented, and it is not yet clear whether the layout is a test or a full rollout.

A knowledge panel is built from structured facts a company can claim and ask Google to correct. An AI Overview is written each time from whatever sources Google retrieves, and on brand queries those sources are mostly someone else's. In Ahrefs' 100-brand panel (see Number of the Month), Wikipedia was cited in 71.9% of the brand-name overviews that showed a source list, YouTube in 38.6% and LinkedIn in 22.8%.

AI assistants answer local questions from official sources, not local news

Searchable, which sells AI search visibility software, ran about 30,000 prompts about 100 UK towns through ChatGPT, Gemini, Perplexity, AI Mode and AI Overviews, collecting 244,725 citations from 10,000 sites. ChatGPT drew on council or government sources for 81.3% of local questions and on local news for 0.9%. On crime questions, police websites appeared in 95.9% of answers.

ChatGPT's background searches pulled up 12.3 sites per answer on average, and it used about four of them. On factual local questions, ChatGPT reached for the kind of official page that Google's Oct. 15 test would put at the top.

Texas sues OpenAI and lists public humiliation among the harms

Texas Attorney General Ken Paxton sued OpenAI and CEO Sam Altman in Collin County district court under the state's Deceptive Trade Practices Act, his office announced on Oct. 23. The release says chatbots without adequate safeguards "can output inaccurate or harmful information, without this risk being disclosed to users," and counts among the harms that "professionals have suffered public humiliation." The suit seeks an injunction that would keep minors off ChatGPT and require court-appointed evaluators to approve new models.

Texas follows eight states that sued OpenAI on Oct. 14, alleging the company misled users about ChatGPT's safety, and a Florida suit filed in June. OpenAI said it has worked with experts on safeguards and wants AI rules that apply to the whole industry. The Texas suit puts the accuracy of a chatbot's output before a court as a consumer-protection question.

STUDY OF THE MONTH

What gets a page cited by ChatGPT: the search it runs counts more than the page

A paper posted to arXiv on Oct. 10 studied 158,099 AI answers and 468,037 citations gathered over 90 days of commercial questions in two markets. ChatGPT cites only pages its own search retrieves, the authors found, and it builds those searches from sub-queries it rewrites for each answer, which rarely match what Google or Bing return. That retrieval step explained 32% of the variation in citations; the properties of the pages explained 2.8%.

Wording still counts within a site: pages whose titles and paragraphs closely matched the questions people ask were cited more. FAQ blocks, structured data, summary boxes and page speed showed an effect only when comparing one website with another. ChatGPT also named brands it already knew, sometimes with no supporting page, and in controlled tests a brand name alone reordered its product rankings when the facts were identical.

For schema and FAQ markup, the finding adds a qualifier to earlier advice: those features appear to help a site compete with other sites, while the wording of each page decides which page on that site gets picked.

NUMBER OF THE MONTH
83%
Share of U.S. desktop Google results for branded keywords that carried an AI Overview on Sept. 29, 2026, up from a 61.3% average in July. In a separate panel of 100 bare brand names, overviews went from 5 of 98 results on July 1 to 63 of 100 by Sept. 30.
FROM STATUS LABS

What do I do if Google shows negative content at the top of my name search?

Status Labs' Oct. 16 guide sets out a seven-step plan for a name search that leads with bad news, from a private-window baseline to monthly re-checks. Its section on AI summaries advises asking the major AI tools what they say about a name, then checking every page they cite as carefully as any other result in the name search.

Can a reputation management agency remove Wikipedia content about me?

The Oct. 20 post explains why no agency can simply delete Wikipedia content, then walks through the routes that exist, from a disclosed talk-page request to the noticeboard for biographies of living persons. The post warns: "A mistake in an article can therefore surface well beyond Wikipedia."

ONE MORE THING

OpenAI's newest model gets fewer answers wrong, by OpenAI's count

OpenAI says GPT-6.1 Terra, released Oct. 26 for ChatGPT Work and Codex, cut its error rate on an internal set of error-inducing prompts from 53.6% to 16.7% at low reasoning effort, and to 4.3% at maximum effort. The test is OpenAI's own, and Terra does not power the everyday Chat tab. Even on those figures, one in six answers to those prompts at low effort still came back wrong.