ChatGPT Shopping Optimization

ChatGPT Shopping is the product surface OpenAI ships inside ChatGPT — a carousel of product recommendations that appears in response to shopping-flavored prompts, plus inline product mentions in answers. Launched in beta and expanded through 2024-2026. It pulls product data from a combination of sources: Google Merchant Center feeds, direct merchant feeds submitted via OpenAI's merchant program, and on-page Product/Offer/AggregateRating schema fetched by OAI-SearchBot. Companion to ChatGPT Instant Checkout, the in-chat purchase flow that went general availability in the US in February 2026.

• Shopify and Etsy: already integrated — your catalog is auto-ingested with no application or feed setup required.
• Other US merchants: apply at chatgpt.com/merchants and submit a direct product feed (CSV, TSV, XML, or JSON; refresh as often as every 15 minutes).
• Google Merchant Center feeds: ChatGPT also ingests via GMC for merchants who've set that up for Google Shopping — one feed, two surfaces.
• On-page schema: Product/Offer/AggregateRating markup on PDPs is read by OAI-SearchBot and used as additional signal. Required for citation in answer text even on products that aren't in the carousel.

ChatGPT Instant Checkout is the in-chat purchase flow that lets users complete a transaction without leaving ChatGPT. The user's payment method, shipping address, and consent are handled inside ChatGPT; the merchant receives the order through the Agentic Commerce Protocol (ACP) — OpenAI's open protocol for agent-mediated transactions (the underlying spec is co-developed with Stripe). Went general availability in the US in February 2026 for ChatGPT Plus, Pro, and Free users. Shopify and Etsy catalogs are auto-integrated; other US merchants apply at chatgpt.com/merchants and implement ACP.

Agentic Commerce Protocol (ACP) is the open protocol OpenAI and Stripe co-developed to standardize the agent-to-merchant transaction flow. Sits at the same layer as Google's Agent Payments Protocol (AP2) — both define how AI agents authenticate as a user's delegate, present cryptographically signed mandates, and complete payments. Merchants typically need to implement both ACP and AP2 to maintain compatibility across the ChatGPT and Google AI surfaces through 2026. Full protocol detail on /agentic-commerce.

• Price: in Offer.price and visible on the PDP, with consistent currency formatting.
• Availability: Offer.availability (InStock, OutOfStock, PreOrder) — ChatGPT actively suppresses out-of-stock products from carousel.
• Reviews and rating: AggregateRating with reviewCount and ratingValue. The single biggest discriminator between near-identical SKUs.
• Brand and MPN/GTIN: Product.brand and Offer.gtin / mpn. ChatGPT uses these to group variants and de-duplicate identical products across sellers.
• Description and category: the natural-language text ChatGPT lifts when describing the product inline in an answer.
• Shipping cost and time: Offer.shippingDetails — increasingly weighted on comparison prompts (“cheapest X with free shipping”).

The carousel is one surface; the answer text is another. Carousel placement depends mostly on feed health and review depth. Citation in the answer text — where ChatGPT writes “the X model from Brand Y is well-reviewed for Z” with a link — depends on the same signals as ChatGPT Search: citation-grade passage structure, OAI-SearchBot access, schema fidelity, and source authority. The right play is to optimize both. Carousel placement drives the direct-to-purchase path; answer-text citation drives the “tell me which to buy” advisory path that often converts at higher AOV.

Discrimination factors, in roughly the priority order we observe: (1) availability and price — ChatGPT favors in-stock products at competitive prices; (2) review depth and rating — AggregateRating-backed signal heavily weighted; (3) shipping cost and time — free shipping or fast shipping flags surfaces products on comparison prompts; (4) seller authority — Organization schema, brand entity authority, and prior citation history; (5) feed freshness — products in feeds that refresh frequently signal active merchants. A merchant who's second-best on every factor still loses to the merchant who's top-3 on most factors. Excellence on review depth and feed freshness is the asymmetric advantage most merchants underweight.

No — ChatGPT Shopping surfaces products whether or not Instant Checkout is wired up. Without Instant Checkout, the user clicks through to your site to purchase; with Instant Checkout, the user completes the purchase inside ChatGPT. Practical implication: ship the carousel-readiness work (feed, schema, reviews) first; layer Instant Checkout / ACP on top once the carousel surface is performing, or where the category warrants the implementation lift. The ACP work is on /agentic-commerce.

Shopify Plus: Shopify is auto-integrated with ChatGPT Shopping — catalog ingestion happens through Shopify's native integration with no merchant action required for carousel inclusion. Instant Checkout work is also natively supported. Easiest path. BigCommerce: requires direct feed submission at chatgpt.com/merchants and ACP integration for Instant Checkout. We've been an Elite BigCommerce Partner since 2012; the integration is standard work. Adobe Commerce / Magento: direct feed submission and ACP integration via our 15-year Magento/Adobe Commerce team. Platform-specific feed depth lives on /optimize-ecommerce-site-ai.

• Shopping prompt panels: 50–200 category-relevant shopping prompts run weekly through ChatGPT, logged for carousel presence, inline citation, and competitor mentions.
• Merchant Center / feed diagnostics: weekly review of disapprovals, warnings, and feed-freshness metrics — leading indicator of carousel demotion.
• Schema validator: automated check that Product/Offer/AggregateRating on every PDP matches the feed source-of-truth.
• Referral and Instant Checkout reporting: GA4 segment for chatgpt.com referrals, plus OpenAI's merchant-side reporting on Instant Checkout order volume where it applies.