Agent Beat for October 22, 2026
Today on Agent Beat: While Amazon blocks AI agents, Walmart rolls out the welcome mat, plus Meta says Muse has over 10M users, with shopping a top activity.
- While Amazon blocks AI agents, Walmart rolls out the welcome mat
- Meta says Muse has over 10M users, with shopping a top activity
- ChatGPT ads change marketing: reading intent from the conversation instead of audience attributes
- ChatGPT, Claude, Perplexity and Google AI Mode: why their citations cannot be compared
- ChatGPT Ads expands to 26 new countries as over 100,000 advertisers have used the platform
- How we built ChatGPT’s novel ad stack for user trust and relevance
- China Wrote The First GEO Rulebook After A TV Scandal: Western SEOs Should Take Notes
- Generative AI: for the same buyer, four engines describe four different markets
- AI Overviews have nearly doubled across five commercial search modifiers
- GEO myths put to a data check: what 96,418 URLs reveal about AI citations
- AI Share of Voice: What It Is, How to Measure It, and What It Actually Tells You
- Google Removes AI Overviews Labels On Some AI Overviews
Transcript
Click any sentence to play from there.
Welcome to Agent Beat. It's Thursday, October 22, 2026. I'm your host for today's briefing. This is the daily briefing on how AI finds brands and how agents choose what to buy. We'll follow the day's developments in retail, advertising, search and measurement.
Walmart is rolling out the welcome mat for AI shopping agents. The retailer is introducing technology that lets AI agents make purchases on a customer's behalf, with controls over spending and approval of transactions, TechCrunch reports. Amazon is taking the other position, and blocks AI agents from shopping on its site. The two retailers are taking opposite positions as agents start to place orders. The headline puts the contrast plainly: while Amazon blocks AI agents, Walmart rolls out the welcome mat.
Meta also gave an update on its own AI app. Meta says its Muse app has over 10 million users, with shopping a top activity. On its Q3 earnings call, Meta said the app has more than 10 million users and that shopping is one of the top activities, TechCrunch reports. The report says users are also adopting the app for personal productivity.
In advertising, ChatGPT ads are changing marketing by reading intent from the conversation instead of audience attributes. Digiday Japan reports that OpenAI is expanding the ChatGPT ads business, which the company says has reached $1 billion in annualized revenue. The article says the ads target intent read from the conversation rather than audience attributes, and that OpenAI frames them as a way to keep the free version available.
Citations are harder to compare than they look. Journal du Net ran 25 French queries three times in each of four AI environments: ChatGPT, Claude, Perplexity and Google AI Mode. It found wide differences in which sources the tools cite. The article argues that a brand has to know what a GEO measurement captures before steering strategy by it. The headline says their citations cannot be compared.
OpenAI's ad platform is also expanding geographically. OpenAI says ChatGPT Ads is now open in 26 more countries and that more than 100,000 advertisers have used the platform. The company says the ads reach people as they explore, compare and decide what to buy inside ChatGPT.
OpenAI also described the ad system behind ChatGPT Ads. The company calls it a novel ad stack for user trust and relevance. In a post, OpenAI says a model reads conversational intent, and that ads are kept separate from the answers. The post says the stack is built to serve relevant ads at scale.
China has published a five-part GEO standard. Search Engine Journal says it is the first GEO rulebook and that it gives Western SEO teams a framework for measuring AI visibility, judging vendors and separating optimization from manipulation. The article links the standard to a television scandal. The headline says Western SEOs should take notes.
For the same buyer, four engines can describe four different markets. Journal du Net put the same buyer questions to four AI assistants three times each and found the leading brand changed from one engine to the next. The piece says being named in an answer is not the same as being recommended.
AI Overviews have nearly doubled across five commercial search modifiers. Semrush says AI Overviews now appear on 88% of the commercial keywords it tracked, up from 46%. The study breaks the change out by reviews, pricing and comparison modifiers. The comparison is 88% now and 46% before.
SISTRIX put common GEO advice to a data check. It compared 48,527 pages cited by ChatGPT, Google AI Mode or AI Overviews with uncited pages from the same domains. The pages were drawn from 96,418 URLs in its AI index. The company says small samples mislead: the measured effect of FAQ schema reversed as its sample grew. It lists template items, such as a table of contents, Article schema, an author box with a photo and a visible update date. It also lists per-page items, such as an FAQ block with schema, a TL;DR and a question used as an H2. The lists cover template items and per-page items, according to the company.
Ahrefs explains AI share of voice, a measure of how visible a brand is in AI answers compared with competitors. The post explains what AI share of voice is, how to measure it, and what it actually tells you. It also sets out what the number does and does not show.
Google is removing the AI Overviews label from some AI Overviews results, Search Engine Roundtable reports. The label is missing on desktop while the mobile version still shows it.
That's today's Agent Beat. Thanks for listening. For the full written brief, with more detail on every story, go to agent-beat.com. Thanks again for making Agent Beat part of your day. We'll be back tomorrow.