SEO Recap covering October 2, 2026: Shopify AI channel auto-enrollment and Google content guidance updates
Daily summary of what matters in SEO, GEO, AEO, and AI search generated with Claude Code (beware of hallucinations)
Generative Engine Optimization (GEO) & Answer Engine Optimization (AEO)
Shopify auto-enrolls eligible stores in every new AI shopping channel
Source: Search Engine Journal
- Shopify's "Allow Shopify to manage for me" setting, on by default under Sales channels > Agentic, auto-enrolls eligible stores in each new agentic storefront channel it adds.
- Products reach AI apps through Shopify Catalog (Google reads from Merchant Center). Meta's Muse agent got access by default on Sept 8, 2026, and Amazon blocked Muse on Sept 20.
- Direct checkout lets shoppers pay inside the chat on Meta, Google, and Copilot without visiting the store. ChatGPT instead sends them to the store's checkout.
- When customers pay in-app, Google Analytics and custom pixels don't fire (only server-side tracking does), so filter the Shopify orders list by channel to see what each AI checkout actually sold.
- No channel charges a fee today, but Zuckerberg said Meta expects to take a small cut of transactions over time. AI-referred sessions to Shopify storefronts grew 197% (~3x) year over year in Q2.
AI in Search / AI Overviews
Karpathy recommends controlled English and richer formats for LLM output
Source: Search Engine Journal
- Andrej Karpathy said on X that asking models to explain things in ASD-STE100, a controlled Simplified Technical English with ~900 approved words, produces output he finds more readable.
- He sometimes softens the request to "80% of the way to ASD-STE100" because the standard is stringent, and its own FAQ says it isn't meant for general writing.
- He suggested other output formats as "even better":
- diagrams that are easier to parse
- "in HTML" for interactive web pages
- "3b1b style" explainer videos narrated via an ElevenLabs API key or a free local alternative
- A reply linked an open-source GitHub skill that applies STE rules to text AI agents read, though it doesn't include the standard's dictionary and isn't built for marketing copy.
Building an SEO content roadmap for the AI search era
Source: Search Engine Journal
- Anchor every keyword, topic, and cluster to a business outcome such as a lead or revenue, not just SERP visibility or traffic.
- Ahrefs found the top-ranking page loses ~58% of its clicks when an AI Overview is present, so read the full SERP before committing to a term.
- Build a keyword universe via Content Gap analysis, cluster by Parent Topic, and map each cluster to a specific hub, service, or article page.
- Prioritize on difficulty, potential, intent, and AI Overview reality, and flag terms where the goal is being cited in the AI Overview rather than earning a click.
Google AI Mode monitoring and OpenAI dots make search requests persistent
Source: Search Engine Journal
- Google opened AI Mode info monitoring to all users on Sept 28 (previously Ultra and Pro only), and OpenAI launched its always-on "dots" agents on Sept 29.
- Both keep a request active after the user leaves. Adding "keep me updated" to a search creates a Google monitoring task that scans web content plus Google's real-time finance, shopping, and sports data.
- Dots use connected apps, a browser, and a cloud computer to handle ongoing work, returning drafts, analyses, and proposed actions for review.
- Later updates can surface information that did not exist when the request was made, including new local businesses, back-in-stock and price-drop alerts, and newly published articles.
- Search Console counts AI Mode impressions and external-link clicks under the Web search type, but as of Sept 30 neither the AI features page nor the reports mention monitoring updates.
AI Overview link tracking parameters turn out to be a browser extension
Source: Search Engine Roundtable
- Tom Pool spotted a "?st_source=ai_overview" parameter appended to links in AI Overviews and AI Mode, raising hopes it signaled proper attribution reporting.
- It turned out to be a Chrome browser extension modifying the URLs, not Google, presumably to sell the data to brokers.
- Barry Schwartz was unable to replicate the parameter across several browsers.
Technical SEO
Google updates people-first content guidance on fake authors and main content
Sources: Search Engine Journal · Search Engine Roundtable
- Google updated its "creating helpful, reliable, people-first content" page on Oct 1, the same day it revised its AI-generated content guidance.
- A new paragraph calls fabricated creator profiles a form of deception and "a signal of a low-quality page":
- AI-generated headshots
- made-up names
- false credentials posing content as human expert work
- A new section defines "main content" as any part that helps the page achieve its purpose:
- primary text and media
- interactive features like calculators and tools
- user-generated contributions
- tabbed or expanded sections
- page titles and headings
- Google added the attributes raters use to judge main content quality:
- effort (human work, not bulk AI output without oversight)
- originality
- talent or skill
- accuracy, which must be high and match expert consensus for YMYL topics
- The guidance mirrors the Search Quality Rater Guidelines but doesn't explain how automated systems would detect a fabricated profile. Search Liaison said in January 2024 that bylines don't help pages rank.
Organic Search & Algorithm Updates
Bing tests a "Recommended by" product results section
Source: Search Engine Roundtable
- Microsoft Bing is testing a "Recommended by [company]" section in product results, spotted by Sachin Patel.
- Product thumbnails link to retailers rather than the recommending source, and there is no link or way to verify that the brand actually made the recommendation.
- Google tested a similar "recommended by" product section back in May 2025.