Someone Just Asked ChatGPT What the Best Product in Your Category Is. Your Product Didn't Appear. Your Competitor Did.

AI product recommendation engines read structured data first — schema markup that specifies price, availability, materials, certifications, and product attributes in machine-readable format. They then cross-reference brand entity verification (Wikidata, Knowledge Graph) before deciding whether the brand is trustworthy enough to cite. Finally, they weight multimodal content — text, image, and video on the same topic — more heavily than text-only content. Most e-commerce brands fail the first two checks entirely, which is why they don't appear regardless of how well their SEO performs.

Yes — the schema architecture is built to scale across large catalogues. We structure product schema at the category and product-type level, not just per individual SKU, which means the structured data framework covers your entire catalogue efficiently. The hallucination audit and SoLLM testing focuses on your highest-value product categories first, then expands. For very large catalogues, we scope the engagement during onboarding to prioritise the product categories where AI visibility will have the most commercial impact.

Seasonal product changes are exactly what Phase 10 (monthly compounding) handles. Every month, we update schema for new product lines, revised pricing, and stock changes — and run a freshness pass to ensure AI platforms are reading the latest data rather than cached old information. For seasonal businesses, the monthly cadence also includes a pre-season schema update before each major selling period to ensure your new products are AI-readable from day one of the season, not weeks into it.

Image Verbalization is the process of rewriting every product image's alt text and associated metadata to be semantically rich and AI-readable. Most e-commerce product images have alt text like "product-img-047.jpg" or "Blue sweater" — which tells AI nothing useful. Functional Alt-Text describes the product in the way a buyer would search for it: "100% RWS-certified merino wool crew-neck sweater, forest green, sizes XS–XXL, available now." This makes your products discoverable in Google Lens AI Mode and visual AI search — the fastest-growing product discovery modality, and one that most e-commerce brands are currently entirely absent from.

The infrastructure we build is primarily for your own domain — the schema, entity verification, FAQ content, and prompt-specific pages are all built on your site. However, the entity authority work (Wikidata, Knowledge Graph, and third-party review platforms) benefits your brand across all platforms, including Amazon. AI product recommendations often cite brands rather than specific retailer listings — so a brand with strong entity authority will be recommended even when the buyer then purchases via Amazon. The SoLLM Baseline Report tracks brand-level citation across platforms, giving you visibility into how AI mentions affect your overall sales across channels.

Sustainability credentials are one of the highest-value schema attributes for AI product recommendations in most categories — because buyers increasingly specify them in queries ("sustainable", "ethical", "certified organic", "recycled"). We structure your certifications — B Corp, RWS, GOTS, Fair Trade, recycled content percentages, carbon offset credentials — as verified, machine-readable schema so AI can cite them accurately rather than generalising. Sustainability hallucinations (AI claiming or denying certifications incorrectly) are also covered in the Phase 8 audit, since these claims carry both commercial and reputational weight.

The SoLLM Baseline Report gives you a concrete before-and-after citation rate: "You appear in X/30 product research prompts before the build. You appear in Y/30 after." Phase 9 then configures GA4 to surface AI-referred traffic as its own channel — so sessions from ChatGPT, Perplexity, and other AI platforms show up separately rather than as "direct" traffic. Over time, you can track AI-referred sessions, revenue attributed to those sessions, and the correlation between your SoLLM score and overall brand search volume. For most clients, the revenue attribution from previously invisible AI-referred traffic is the most surprising result — it was always there, just untracked.

Shopify and WooCommerce SEO plugins optimise for Google's traditional search crawler — which reads pages differently from AI systems. AI crawlers (GPTBot, AnthropicBot, PerplexityBot) have different access requirements, different schema interpretation, and different entity verification processes. The plugins don't configure llms.txt, don't build entity authority on Wikidata or Knowledge Graph, don't write Functional Alt-Text for visual AI, and don't track SoLLM citation rates. They're built for a search world that's rapidly becoming less relevant for product discovery. We build for the AI product discovery layer — which those tools don't address.

Yes — and the two disciplines genuinely don't conflict. Your SEO agency optimises for Google's traditional search crawler and human-click metrics. We optimise for AI citation infrastructure and machine trust signals. The schema we implement is additive to what an SEO agency does, not duplicative or contradictory. The entity work we do (Wikidata, Knowledge Graph, review platform consistency) typically improves traditional SEO performance as a side effect — but that's not the primary objective. Many of our clients run both simultaneously without any conflict.