This week in search October 23, 2026

OpenAI expanded ChatGPT ads to 26 more countries, bringing the total to 89, and says over 100,000 advertisers have run campaigns in a little over eight months. Ads appear on the Free and Go plans, and Plus, Pro and Enterprise stay ad-free. OpenAI says more than 1.2 billion people use ChatGPT each week.

Semrush tracked 400 commercial keywords in banking, ecommerce and software for a year and saw AI Overview coverage climb from 45.75 percent to 88 percent. “Best X for Y” searches sit highest at 96 percent, and review searches rose fastest. These are the searches closest to a purchase, and few of them come without an AI summary now.

Barry Schwartz rounded up three studies. Marie Haynes found ChatGPT sessions were 0.62 percent of Google organic across 100 sites, though those visitors converted better. Kevin Indig couldn’t detect hidden AI traffic in direct visits. The SEO Works found organic sent 39 times the traffic of AI. Taken together, Google is where the volume is.

Google is testing ads in AI Mode for people who connect their apps through Personal Intelligence, after saying in March there were none. Its help page says ads won’t be based directly on Gmail or Photos, though summaries and inferences from those apps may be used to show them. Google told Search Engine Land the test reverses its earlier ad-free approach.

Walmart announced Agent Profiles, launching in the next month, which give an AI agent its own profile inside a customer’s account with approval rules or a spending limit. Agents see the same prices people do and never receive card details. A gateway for agents is in the works, starting with a beta for companies and developers.

Meta’s new Commercial Terms let it crawl sites tied to businesses using its products, and an early version included training its products on that content. After Nieman Lab’s reporting, Meta removed training from the clause and kept the crawling rights. A spokesperson says the crawling terms only cover ads and business products, and legal experts say the wording is still broad.

Gary Illyes explained that robots.txt controls access, and it doesn’t dictate what a system does with data it has already crawled. Illyes also co-wrote a draft with Mozilla’s Martin Thomson that adds a Content-Usage rule to the file. It’s a working document for now, and Google ignores lines it doesn’t support. The AI agent readiness guide lists which crawlers follow the file today.

Brodie Clark spotted AI-written summaries in the citation cards next to AI Overviews, and he says the text doesn’t appear exactly that way on the page. Barry Schwartz saw it on some of his test queries. Schwartz says Google seems to be testing or rolling it out.

PHP released emergency fixes for two critical flaws, and Google reminded site owners taking servers offline to return a 503. That status keeps indexed URLs in Search while you work. Google warned against generic 404s and error pages served as 200s, and its documentation says to keep a 503 to one or two days at most.