On May 15, 2026, Google published its official Generative AI Optimization Guide. It restates that traditional SEO remains the foundation for AI Overviews and AI Mode, and pushes back on tactics like llms.txt, content chunking, and AEO/GEO hacks. The guide is useful and worth reading. But Google speaks only for Google. ChatGPT, Perplexity, and Claude do not share the same architecture, and several of the things Google calls optional are signals those engines actively use. Here is what the guide says, what it means for Shopify merchants, and what we recommend doing with it.
What Google's guide actually says
The guide, published in the Google Search Central documentation under the title "Optimizing your website for generative AI features on Google Search," focuses on two specific surfaces: AI Overviews and AI Mode. Both are Google products. Google explicitly notes that these features are "rooted in our core Search ranking and quality systems," which is why their guidance is essentially traditional SEO refined for AI synthesis.
The key positions Google takes:
SEO fundamentals still apply
Indexable pages, crawlable content, clean technical structure, valuable non-commodity content, good page experience. None of this is new, and Google reiterates that AI features depend on the same ranking systems as regular Search.
Non-commodity content over commodity content
Google contrasts "7 Tips for First-Time Homebuyers" (commodity, common knowledge) with "Why We Waived the Inspection & Saved Money: A Look Inside the Sewer Line" (non-commodity, experience-based). The framing is useful and worth internalizing.
RAG and query fan-out are the mechanisms
Google describes how its AI features retrieve content (Retrieval-Augmented Generation) and how a single query gets expanded into many related sub-queries to surface comprehensive answers. This part of the guide is technically accurate and useful context.
Pushback on AEO and GEO terminology
Google states: "From Google Search's perspective, optimizing for generative AI search is optimizing for the search experience, and thus still SEO." In other words, they prefer the existing SEO label.
What Google says is not necessary
The guide explicitly names tactics it considers unhelpful for Google's AI features: llms.txt files, chunking content for AI, rewriting content for AI systems, seeking inauthentic mentions, and overfocusing on structured data.
The part of the guide that is genuinely useful
Before getting to what's missing, credit where it's due. Three sections of Google's guide are excellent and directly applicable to Shopify merchants:
Non-commodity content
The commodity-vs-non-commodity distinction is sharp. A Shopify store publishing "10 Reasons to Choose Organic Cotton" is commodity. A store publishing "What We Learned After Pulling 2,000 Returns on Our Cotton Tees" is non-commodity. The second one is what AI engines cite, full stop, including ChatGPT and Perplexity.
Browser agents and DOM access
Google explicitly mentions that browser agents "may access your website to gather the data they need to complete these tasks, such as analyzing visual renderings (like screenshots), inspecting the DOM structure, and interpreting the accessibility tree." This is a critical signal. If your schema markup or product information is JavaScript-injected, browser agents may not see it.
Universal Commerce Protocol (UCP)
The guide cites UCP as an emerging protocol that will allow search agents to perform commerce actions. This is the first time Google has explicitly named UCP in a public guide. Shopify merchants paying attention should bookmark this — it will matter in the next 12 months.
Where the guide quietly speaks only for Google
Now the part that matters. Google's guide is accurate for Google's surfaces. It is not a description of how the broader AI search ecosystem works.
Everywhere the guide says "AI search," the unstated qualifier is "on Google's surfaces." Google has no authority over how ChatGPT, Perplexity, Claude, or other AI engines crawl, parse, and cite content. Treating Google's guide as universal advice for all AI engines is a category error.
Take llms.txt as an example
Google says you don't need llms.txt files for AI Overviews. Fine. That is consistent with what we observe on Google's surfaces. But:
- ChatGPT. OpenAI has publicly stated that GPTBot and OAI-SearchBot consider llms.txt-style endpoints when discovering site context. We've measured citation lift on stores that added a custom llms.txt with structured catalog summaries.
- Perplexity. Perplexity's crawler documentation references AI sitemaps and explicitly recommends machine-readable endpoints for sites that want clearer indexing.
- Anthropic (Claude). ClaudeBot honors structured AI discovery files when present. Claude is increasingly used inside developer agents that read llms.txt to ground their answers.
So when Google says "you don't need llms.txt," the honest reading is: "you don't need llms.txt for our specific implementation." The file is not redundant for the rest of the AI ecosystem.
The same logic applies to structured data
Google's guide says structured data "isn't required" for generative AI search. They also say, in the same paragraph, that it's "a good idea to continue using it as part of your overall SEO strategy." Both statements are true at once. What the guide doesn't say:
- Product schema is heavily used by Perplexity Shopping and ChatGPT product cards.
- FAQPage schema is one of the highest-correlation signals for Answer Capsule extraction across all major engines.
- Organization schema with sameAs links is how AI engines disambiguate your brand from competitors with similar names.
Structured data is not optional infrastructure for a Shopify store serious about AI visibility in 2026. It is foundational.
Our reading: align with Google AND ship the rest
We are not anti-Google. We agree with most of what the guide says. The non-commodity content principle is correct. SEO fundamentals remain the foundation. Page experience matters. Browser-agent-readable DOM is critical.
Where we differ is what merchants should actually do.
Do everything Google's guide recommends
Server-side rendered HTML. Crawlable, indexable pages. Clean technical structure. Non-commodity content. Good page experience. All of it. The fundamentals work for Google AI and for every other engine. There's no trade-off here.
Add the layer Google didn't mention
llms.txt for ChatGPT and Perplexity. Schema-Stitch with linked entities (Organization, WebSite, Product, Article, FAQPage, BreadcrumbList) in one @graph. Answer Capsules in metafields. AI crawler whitelist in robots.txt. These don't conflict with Google's guide. They extend it.
Test on every major engine, not just Google
If your only test is "do I appear in Google AI Overviews," you're optimizing for one of five major AI surfaces. Run prompts on ChatGPT, Perplexity, Gemini, Claude, and Copilot. Measure citation lift across all of them. That's the actual job.
What Shopify merchants should actually do this week
Read Google's guide and apply the SEO fundamentals
It's a good guide. The technical advice is correct. The non-commodity content framing is sharp. If you haven't done the basics, do them.
Verify your schemas render server-side
Use Chrome DevTools or the Rich Results Test. View Source on your product page. If your schema only appears after JavaScript executes, browser agents and several AI crawlers won't see it. This is where most schema apps quietly fail.
Add llms.txt for the engines that read it
Google doesn't need it. ChatGPT, Perplexity, and Claude benefit from it. The cost of adding one is low. The cost of not adding one is being invisible on three engines that already drive real traffic to Shopify stores.
Whitelist AI crawlers in robots.txt
GPTBot, ClaudeBot, PerplexityBot, OAI-SearchBot, Google-Extended, and the rest. Explicit allow beats default. Stores using Cloudflare in front of Shopify should double-check their Cloudflare bot management settings while they're at it.
Audit your content for commodity vs non-commodity
Google's framing is correct. Re-read your top 10 pages and ask honestly: would AI engines cite this because it says something unique, or skip it because it's interchangeable with 1,000 similar pages?
Google's guide is not wrong. It's incomplete by design — Google is publishing what works for Google's products, not a universal AI search optimization playbook. The merchants who win in 2026 will follow Google's advice AND go beyond it. The merchants who treat the guide as the full answer will be invisible on three engines (ChatGPT, Perplexity, Claude) that are already shaping how customers discover Shopify stores.
What this changes for GetCitedShop
Nothing, honestly. Our approach was already aligned with what Google's guide recommends — server-side rendered schemas, clean technical structure, non-commodity content, good page experience. The pieces Google says aren't required (llms.txt, Schema-Stitch, AI crawler whitelist) remain valuable for non-Google engines, and that's where they live in our products.
If you want the architecture that covers Google AI and the rest of the AI ecosystem, the AI-Ready Schema Kits ($79–$599) or the Cited Theme ($299 pre-order) ship with everything pre-configured. If you want hands-on implementation, book a Strategy Call.
Either way, read Google's guide. It's worth your time. Just don't stop there.
— Jonathan
Google's official guide: Optimizing your website for generative AI features on Google Search (updated May 15, 2026).