Google just published a developer-facing playbook for store owners on being visible in Google rich results, Google AI Overviews, ChatGPT, Claude, Perplexity, Gemini, and the next wave of AI shopping assistants. It's the most authoritative single document on what AI search rewards — and it's also one of the most under-read documents in ecommerce right now, because it's filed under a developer URL and written for engineers.
We audited every feature in StoreCue against it. The good news for our merchants: most of what Google asks for is already shipped. The better news: we closed four small distances between "doing the right thing" and "doing the right thing perfectly," and pushed live yesterday. This post walks through exactly what changed, what we deliberately don't do, and why.
What Google said
The guide makes one point loud and clear up front: SEO fundamentals are still the foundation. AI features build on Google's existing ranking and quality systems — they don't replace them. Quote:
"The best practices for SEO continue to be relevant because AI features rely on core Search ranking and quality systems."
From there, the guide drills into eight concrete areas: unique non-commodity content, structured data, semantic HTML, page indexability, original first-hand perspectives, language that matches real shopper intent, trust signals (E-E-A-T), and avoiding duplicate content. It also explicitly debunks four tactics making the rounds in AI-SEO circles right now:
- No "llms.txt" or other AI-specific files. Google doesn't read them. Neither do the AI search engines.
- No chunking content for AI. Google's systems understand the nuance of multiple topics on a single page.
- No keyword-variation harvesting. "Stop worrying about capturing every variation of how someone might search."
- No manufactured "mentions" or link schemes. Authentic perspective beats artificial visibility signals.
How StoreCue maps to each recommendation
What follows is the eight-item map — each Google recommendation alongside the specific StoreCue feature that delivers on it.
1. Unique, non-commodity content
Google: "Develop content with a unique point of view rather than recycling existing information. Focus on non-commodity content that's helpful, reliable, and people-first."
StoreCue's entire scan pipeline is grounded in your Brand context — your voice, your use cases, your restricted phrases, your trust signals. Generic questions like "is this comfortable?" are explicitly forbidden in our scan prompt. Every published answer is anchored to something specific about your product, drafted in your brand's voice, and reviewed by you before it goes live.
2. Structured data (FAQPage JSON-LD)
Google: "Structured data isn't required, but helps make your site eligible for rich results."
Every published Q&A ships as schema.org/FAQPage JSON-LD on the product page — exactly the format Google's rich results documentation requires. The same standard is read by ChatGPT, Claude, Perplexity, and Gemini. One canonical format, every tool reads it.
3. Semantic HTML
Google: "Use semantic HTML when possible to help us understand the structure of your page."
We enhanced the static FAQ layouts in our storefront block to use real <h3> tags for question text instead of generic paragraphs — so crawlers and assistive tech get the heading hierarchy they expect. Accordion / editorial / brutalist layouts already used semantic <details> elements. Every answer ships in the initial HTML — never lazy-loaded via JavaScript, never trapped inside a popup — so Google can read every Q&A whether or not a shopper clicks expand.
4. Indexability and crawlability
Google: "Your pages must be indexable and eligible to be shown in Search with a snippet."
We extended StoreCue's storefront readiness check beyond "is the FAQ block rendering?" to a fuller bundle: is the page reachable, is there a noindex meta tag or X-Robots-Tag header blocking crawlers, is there valid FAQPage structured data on the page. So the dashboard tells you "Google can read this" with the same confidence it tells you "your FAQ block is installed."
5. First-hand customer perspectives
Google: "Provide first-hand reviews and original perspectives based on personal experience."
This is the most under-utilized signal in ecommerce. The questions your real customers ask are first-hand E-E-A-T (Experience, Expertise, Authority, Trust) gold — far stronger than questions an AI guesses from a PDP. Our new Customer questions surface lets you paste raw support tickets, post-purchase surveys, product reviews, or DMs — and StoreCue extracts the distinct shopper questions out of the blob. You review them and promote the relevant ones to custom FAQs on individual products. Your real customer voice, structured for AI search.
6. Shopper-vocabulary bridge
Google: "Write content for human audiences. Don't try to capture every variation of how someone might search — AI systems understand synonyms and meanings."
On the surface this sounds like it argues against bridging brand language to shopper language. It doesn't. The warning is against keyword stuffing and inauthentic variation harvesting. Our Shopper Language Map is the opposite of that: it teaches StoreCue both your brand vocabulary (the distinctive language you've earned) and the real shopper search vocabulary (what your customers actually Google), then weaves them naturally — never replacing your voice, never forced. A footwear brand that calls their products "hands-free" gets FAQs that show up for both "hands-free shoes" and "slip-on shoes" without sacrificing their distinct voice.
7. Trust signals (E-E-A-T)
Google's quality framework looks for evidence of Experience, Expertise, Authority, and Trustworthiness. Return policies. Brand story. Patented technology. Press and awards.
The Trust signals tab in Brand context captures all eight categories — and StoreCue uses them in answers only when they directly help (returns into fit questions, warranty into durability questions, technology credibility into "does this really work?" questions). Never stuffed where they don't belong. Never invented when you leave a field blank. The result is answers that read like they were written by someone who genuinely knows your brand — because, indirectly, they were.
8. Duplicate content prevention
Google: "Minimize duplicate content to avoid bad user experience and wasted crawl budget."
Variant-heavy catalogs (color/size families that live on a single PDP but are technically separate Shopify products) would naturally generate as many FAQ sets as there are variants — exactly the duplicate content Google penalizes. StoreCue's Product Groups feature solves this at the source: scan and approve once on a primary product, and StoreCue publishes the same FAQ set to every variant in the family. One canonical answer set per PDP.
What we deliberately don't do
Google's guide is just as clear about what to avoid as what to do. We don't run any of these — and that's a feature, not a missing one:
- No llms.txt or "AI manifest" files. We only publish standard schema.org FAQPage. Every major AI search engine reads the standard. Custom files don't help.
- No content chunking for AI. The FAQs we publish are natural Q&A pairs — not artificially fragmented content blocks pretending to be "AI-friendly."
- No keyword variation farming. The Shopper Language Map isn't keyword targeting — it's authentic language bridging, capped at the synonyms your real shoppers actually use.
- No scaled content abuse. Every published FAQ goes through your review queue. AI drafts the candidates; you decide what ships.
What we're watching for next
Google's guide is a living document, and we re-audit StoreCue against it on every major update. Three areas we're actively tracking:
- Agentic experiences. Browser agents that read PDPs on shoppers' behalf are emerging. Because we publish standard FAQPage JSON-LD, our answers are already agent-readable — no rework needed when the next wave ships.
- Visual and video FAQs. Google's AI features can surface images alongside text answers. We're scoping image attachments on individual Q&As (a sizing chart on a fit question, for instance) for a future release.
- Search Console signals. Eventually surfacing which of your published FAQs are appearing in rich results, AI Overviews, or AI-tool answers — so you can see the impact directly inside the StoreCue admin. Roadmap, not next sprint.
The bottom line
AI search isn't one channel. It's a constellation of platforms reading the same product pages and looking for the same fundamentals: authentic content, structured data, semantic HTML, customer-voice perspective, brand-vocabulary authenticity, and trust signals. Google just put that constellation in writing.
StoreCue's job — the thing you're paying us for — is to keep up with what the platforms shaping AI search are actually asking for, and to translate it into product improvements you don't have to manage yourself. The four enhancements we shipped this week are exactly that. The eight-item mapping above is exactly that. The fact that you're reading this post is exactly that.
Want StoreCue on your store? Install from the Shopify App Store →