Field notes on product-level AI visibility.
Posts on ChatGPT citation behavior, Perplexity Shopping data, schema and metafields, measurement, and agentic checkout. Written by the eCommerce Insights research team; updated as the engines change.
Alexa for Shopping: Amazon Moves the Answer Layer Above the Results
Amazon merged Rufus into Alexa for Shopping and moved AI overviews to the top of search results and PDPs. What that changes for brands selling on Amazon.
Platforms · 2026-07-23Webflow MCP 2.0: agents can now edit sites without a Designer session
Webflow MCP 2.0 lets agents run most element, component, style, and variable operations without a Designer session, with workspace permissions enforced.
Agentic · 2026-07-13Hydrogen's WebMCP tools let in-browser AI agents work your storefront
Shopify's Hydrogen preview lists WebMCP tools exposing storefronts to in-browser AI agents for catalog search, browsing and cart. What it changes, what is unknown.
Platforms · 2026-07-06Wix Headless Now Auto-Generates SEO. The Schema Types Are Unspecified
Wix-managed headless projects on Astro now auto-generate SEO metadata, robots.txt and sitemap.xml. The changelog never names a schema type.
Platforms · 2026-07-02Wix released an API to control llms.txt files — here's what it means for AI agent visibility
Wix released an API to manage llms.txt files. Auto-generated defaults expose site capabilities to AI agents through MCP endpoints. What merchants should do now.
Agentic · 2026-06-26Shopify deprecates Storefront MCP cart tools for UCP Cart MCP
Shopify deprecates get_cart and update_cart on Storefront MCP in favour of UCP Cart MCP. New endpoint, new metadata, PUT semantics, and an August 31, 2026 date.
Agentic · 2026-06-22Google's Agentic Resource Discovery Spec Is About Tools, Not Your Product Catalog
Google's new Agentic Resource Discovery spec publishes agent-callable capabilities at a domain-anchored file. It is not a product-data standard — here is the difference.
Agentic · 2026-06-19Shopify's standard storefront events and actions: what changes for agents
Shopify.actions.getCart, updateCart and openCart are now documented on every Liquid storefront, with agents named as callers. What it covers, and what it does not.
Platforms · 2026-06-16Webflow's Data API Now Reads and Writes Page JSON-LD Schema
Webflow's Data API now exposes endpoints for page JSON-LD schema. What changes for ecommerce catalogs, what the beta tag implies, and what the changelog leaves out.
Agentic · 2026-06-12Shopify just made every store agent-readable.
Five agent endpoints are live by default — llms.txt, agents.md, a UCP profile. What the defaults say, and the brand layer worth adding.
Agentic · 2026-06-10Agentic checkout: what brands should do now.
Draft carts are live, ACP and UCP are in pilot. The staged read, and the no-regret work that pays off at every stage.
Platforms · 2026-06-10AI model knowledge cutoff dates: why they matter less than you think.
Vendor-published cutoffs for ChatGPT, Gemini, Claude, Copilot, and Perplexity — and the live-retrieval column that decides product answers.
Platforms · 2026-06-10Artificial intelligence search engines: the list that matters.
Eight AI search engines on one ruled table — live web access, shopping surfaces, citation behavior — with the ecommerce read for each.
Data · 2026-06-10Does ChatGPT give everyone the same answer?
Sampling, memory, retrieval, geography, model routing — five reasons your visibility is a distribution, not a fact, and how to sample it.
Technical · 2026-06-10Robots.txt for AI crawlers: who to allow and why.
Training bots, answer bots, and action agents — what each user-agent controls, plus a starter policy for ecommerce.
Strategy · 2026-04-18Brand-level AI tracking is missing your revenue.
Brand mentions are a comms metric. The SKU is the revenue unit — and brand tracking cannot see which SKUs are silent.
Platforms · 2026-04-18ChatGPT vs Google AI Overviews for D2C.
Who is asking, how each engine cites, what wins in each, and where the first hundred hours of effort go.
Technical · 2026-04-18How AI engines pick which products to cite.
Retrieval, grounding, ranking, and source-trust signals. Mechanism first, no hand-waving.
Data · 2026-04-18Measuring AI visibility: what actually matters.
The per-SKU-per-engine primitive, the rollups that mislead, and the dashboard a merchandiser can act on.
Data · 2026-04-18Perplexity Shopping citation patterns.
Citation counts, source mix, and where D2C PDPs lose to review sites — from a 200-query manual sample.
Technical · 2026-04-18The schema your product pages need for AI answers.
Product JSON-LD field by field, the fields most PDPs leave empty, and a complete working example.
Technical · 2026-04-18Shopify metafields for AI search.
Which metafields matter, how to structure them, and how they flow into Product JSON-LD and llms.txt.
Strategy · 2026-04-18The D2C playbook for AI search, 2026 edition.
Seven opinionated steps ranked by return on effort, honest about the parts of GEO practice still forming.
Strategy · 2026-04-18Why GEO differs from SEO.
Five similarities, five differences, who each serves, and why most teams need both disciplines funded.
Strategy · 2026-04-18Your best seller is invisible in ChatGPT.
A thirty-minute audit: five prompts, a short rubric, and a triage list for the PDPs that need work this month.
Related hubs
Ready to see your catalog?
A product-level visibility report across six AI engines, with concrete PDP fixes for every SKU you carry.