What is ChatGPT shopping?
When a user asks ChatGPT for a recommendation — “best standing desk under $500”, “is this serum good for sensitive skin” — the assistant increasingly responds with a shopping-style answer: named products, key specs, price ranges, pros and cons, and links to buy or read reviews. With roughly 900 million weekly users (OpenAI, 2026), this is now a top-of-funnel product discovery channel that merchants cannot ignore.
How does ChatGPT pick which products to show?
The product pipeline mirrors the general four-stage pipeline:
- Retrieval. Product questions fan out into sub-queries (price, specs, use case, comparisons) resolved against the Bing index via OAI-SearchBot’s index and live page fetches by ChatGPT-User.
- Grounding. Product facts — price, availability, materials, ratings — are cross-checked across your product page, retailer listings, review platforms and editorial coverage. Inconsistencies kill recommendations.
- Generation. The model composes a recommendation narrative, usually with trade-offs.
- Citation. Sources that survived grounding appear as links: product pages, review sites, comparison articles.
Because web search in ChatGPT depends on the Bing index, Bing Webmaster Tools verification and IndexNow submission are table stakes. See why the Bing index decides ChatGPT visibility.
What data do we have?
OpenAI reports roughly 900 million weekly ChatGPT users in 2026, and has repeatedly stated there are no ad slots inside answers as of November 2026 — recommendations are organic. That makes shopping answers a pure GEO play: you cannot buy placement, you can only earn it. The broader context matters too: with roughly 68% of search sessions ending zero-click (Ahrefs, February 2026), the recommendation itself — even without a click — shapes purchase decisions.
How to optimize products for ChatGPT shopping
- Verify your site in Bing Webmaster Tools and enable IndexNow so product and price changes are indexed within hours.
- Ship complete Product structured data: name, brand, GTIN where available, price, availability, aggregateRating, review.
- Keep price and stock identical across your site, marketplaces and feeds — grounding penalizes mismatches. See pricing pages for AI.
- Build review coverage on platforms ChatGPT cites: category review sites, Reddit threads, YouTube comparisons. See review management.
- Publish honest comparison and “best for” content that names your product’s ideal use case — assistants reuse this framing. See comparison content.
- Answer pre-purchase questions (sizing, compatibility, shipping, warranty) in crawlable text, not images or accordions loaded by script alone.
Common pitfalls
Merchants frequently hide prices behind login or “request a quote”, which removes the product from price-sensitive recommendations. Others let marketplace listings contradict the brand site on specs, which grounding treats as unreliable. A third mistake is optimizing only the product page while the assistant actually cites third-party reviews and editorial comparisons — your source graph matters as much as your own page. See source graph building.
How does this differ from Google shopping surfaces?
Google’s AI surfaces lean on Google Business Profile, Merchant Center and the Shopping graph; ChatGPT leans on the Bing index plus live fetches and third-party sentiment. A dual-engine approach is covered in multi-engine strategy, and marketplace-specific guidance in GEO for marketplaces.