ChatGPT Feeds & Ads: everything you need to know
Best for: brands and retailers who want their products recommended in ChatGPT’s shopping answers today, through a structured product feed, and who are preparing for sponsored placements as ChatGPT Ads open market by market.
|
|
|
|
| Fast | Longer |
|
|
|
|
|
|
| Discovery | Purchase |
ChatGPT has become a place where shoppers describe what they need, compare options and decide, in one conversation. For merchants there are two ways in. ChatGPT Shopping surfaces product results (images, prices, ratings, reviews, merchant links) ranked on availability, price, quality and whether you're the maker or primary seller: organic results, powered by a product feed you push to OpenAI, not by ads. ChatGPT Ads adds a paid layer on top: a single sponsored card below the answer, clearly labelled and visually separated, which never changes what ChatGPT actually says.
Both layers run on the same asset. The organic side is stricter than Google Shopping in one respect: there is no bid to compensate for weak data, so a product either surfaces or it doesn't, based on how structured and complete the catalogue is. The paid side targets conversation context rather than keywords, with advertiser-written "context hints" guiding the match, and OpenAI's Ads Manager supports campaigns built directly from a product feed.
Through Lengow, merchants have supported the OpenAI Product Feed since day one: catalogue mapping to the specification, enable_search and enable_checkout flags, updates as often as every 15 minutes, and validation before going live. That same feed now also powers ad campaigns: after piloting in the US from January 2026 and adding the UK in June, OpenAI took ChatGPT Ads live across all 27 EU member states plus Iceland, Liechtenstein, Norway and Switzerland, 31 markets at once, on 24 August 2026, with self-serve Ads Manager access opening across every one of them a week later. European merchants can now build feed-based ChatGPT Ads campaigns through Lengow the same way US and UK advertisers already do.
Key figures
Sources: OpenAI, Ads in ChatGPT: The Basics (pricing and mechanics, 2026); OpenAI, ChatGPT Ads expands across Europe (August 2026 rollout and self-serve dates); Lengow, How to get ready for ChatGPT Ads in Europe; Bain research on GenAI shopping behaviour; Lengow, ChatGPT Product Feed specification. The ad product continues to evolve and its mechanics may change.
Three layers, three speeds
Who will see your products?
The audience shops by conversation: they describe a need ("waterproof hiking boots under €120"), refine it with follow-up questions and compare a shortlist without leaving the chat. Organic shopping results reach every ChatGPT user having a shopping conversation; ads reach only logged-in adults on the Free and Go plans, since Plus, Pro, Business and Enterprise accounts never see them.
The behaviour is moving faster than the transactions: roughly one in eight US consumers now starts a shopping journey on a generative AI chatbot, yet AI-sourced traffic still converts markedly worse than classic search. That gap is why OpenAI itself softened its checkout ambitions, and why, even with the ad business now past a $1bn annualised run rate, the honest framing for this channel is qualified discovery with fast-growing paid reach, not yet a revenue line you can forecast with precision.
The discipline both layers demand is the same: structured, complete, fresh product data, because the model reasons over your feed, and a sponsored card can't rescue a product the organic answer wouldn't have recommended.
Brands & primary sellers
Ranking favours the maker or primary seller of an item, so brands selling their own products start with a structural advantage over resellers.
Attribute-rich catalogues
Products whose data answers natural-language criteria (use cases, materials, compatibility, reviews) match how people actually describe what they want.
European advertisers, right now
With self-serve access now open across the EU, EEA and Switzerland, European merchants can test a young auction while it's still cheap relative to Google and Meta; that pricing won't survive maturity.
A weaker fit: attribution-dependent, ROI-now advertisers
Conversion measurement is still maturing everywhere the channel runs, checkout has already been scaled back once, and AI traffic converts below search. Budget-holders who need precisely forecastable returns this quarter should still treat this as early, fast-moving territory.
Who runs ChatGPT as a channel?
Ownership splits along the channel's two layers: organic visibility sits with the people who own product data and generative engine optimisation, while the ad business belongs to performance teams running discovery budgets.
SEO & GEO leads
Own visibility in generated answers. Their lever is the feed plus brand-controlled sources, and their reality check is simple: search your own best sellers in ChatGPT the way a shopper talks.
Performance marketing leads
Run the ad tests now that self-serve inventory exists across the US, UK and 31 European markets. Their adjustment: context hints replace keywords, budgets are discovery budgets, and measurement runs on a maturing pixel plus UTM parameters.
E-commerce & product data managers
Hold the feed itself: identifiers, availability, prices refreshed up to every 15 minutes, and the enable_search and enable_checkout decisions per SKU and market.
What you need to know before starting
Split the work the way the channel splits: an organic feed open to merchants now, a self-serve ad auction now live across the US, UK and continental Europe, and a checkout protocol worth flagging per SKU without building a roadmap on it.
Merchant profile
Any merchant can apply for direct feed access with OpenAI; Shopify merchants are integrated automatically via Shopify Catalog. Advertisers additionally need Ads Manager access in a live market.
Cost model & budget
Organic results are free. Ads run CPC (Clicks objective) or CPM (Reach objective) in a relevance-weighted second-price auction, with bids set per ad group; OpenAI's original guidance suggested starting around $3-5 maximum CPC, though actual costs vary as the auction matures market by market.
Countries & rollout
The product feed is not geographically gated the way ads are. Ads: the US (self-serve since May 2026), the UK (June 2026), then all 27 EU member states plus Iceland, Liechtenstein, Norway and Switzerland, live to users from 24 August 2026 and self-serve for advertisers from 31 August 2026. Ads still show only on Free and Go plans, and European access runs under GDPR-compliant opt-in consent.
Technical setup
The feed is pushed to a secure OpenAI endpoint (CSV, TSV, XML or JSON), validated before going live, and refreshed on schedule. Instant Checkout requires ACP-compliant endpoints with a verified PSP such as Stripe. Ads run from OpenAI's Ads Manager, self-serve in every live market.
Feed requirements
Complete identifiers, titles, factual descriptions, prices, availability and GTINs, plus reviews and ratings where available, structured attributes over marketing copy, seller name and published return, privacy and terms policies, and the enable_search / enable_checkout flags set deliberately.
Measurement
Ads Manager reports impressions, clicks, spend, CTR, average CPC/CPM and conversions; conversion measurement is still maturing, so pair it with UTM parameters and consented first-party data, and benchmark against the reality that AI traffic converts below classic search.
How to get your products into ChatGPT with Lengow
The work starts with the catalogue: Lengow maps it to the OpenAI Product Feed specification, enriches titles and attributes for natural-language matching, sets the search and checkout flags per segment, and automates updates on the cadence the channel supports. The same structured feed now also becomes the raw material for feed-based ad campaigns in Ads Manager, including for merchants advertising into the newly live European markets.
The channel rewards merchants who treat product data as the ranking factor it now is. Check out Lengow's ChatGPT Product Feed specification guide.
| Channel challenge | What successful merchants do | What Lengow automates |
|---|---|---|
| No bid rescues weak data | Fix structural issues first: duplicate IDs, missing GTINs, inconsistent categories, messy variants. | One canonical catalogue with rules, validation and channel-specific error fixing. |
| Natural-language matching | Write factual, attribute-rich titles and descriptions that map to how shoppers describe needs, not keyword stuffing. | Rules restructure titles, descriptions and attributes per channel without touching the source catalogue. |
| Ranking signals | Keep availability, price and seller status accurate, the factors OpenAI says drive merchant ranking. | Automated updates as often as every 15 minutes keep prices and stock nearly real-time. |
| Search & checkout flags | Set enable_search and enable_checkout deliberately, starting search-only and extending checkout per SKU or market once ACP endpoints are ready. | Flags are managed per segment, so eligibility decisions scale across the catalogue. |
| Reviews & trust signals | Include ratings, reviews, seller name and published return, privacy and terms policies, which the experience factors into comparisons. | Review and policy fields are carried in the feed structure alongside product data. |
| Variant clarity | Consolidate variants with stable URLs so the model reads one product family, not a dozen near-duplicates. | Variation handling structures parent-child data to the specification. |
| Feed-based ad campaigns, now across Europe | Build campaigns from the product feed in Ads Manager, with context hints describing where products fit, now that self-serve access spans the US, UK and 31 European markets. | The same optimised feed serves organic Shopping and feed-based campaigns, one workflow, one catalogue, every live market. |
| Measurement gaps | Add UTM parameters, set up conversion measurement, and build consented first-party data as attribution matures in every newly opened market. | Tracking parameters are managed in the feed URLs by rule. |
| Multi-surface consistency | Keep ChatGPT, Google Shopping and marketplace data consistent, since shoppers cross-check and AI models read everything. | One catalogue distributed to ChatGPT, Google Shopping and 1,800+ other channels. |
With Lengow AI Studio, AI is embedded directly into the feed workflow: AI Category Mapping keeps taxonomy consistent, AI Content Localisation adapts product content per market, a genuine advantage now that campaigns can run across 30+ European markets from one catalogue, and AI Image Compliance prepares visuals at scale, the same catalogue quality conversational surfaces now rank on.
Should ChatGPT be on your roadmap?
The channel has already delivered one lesson in humility: OpenAI itself scaled back native checkout after shoppers browsed far more than they bought. What survived that correction, and then some, is the part that was always solid: a feed-ranked discovery surface used by a very large audience, now paired with a fast-scaling ad business that reached a $1bn annualised run rate within months of opening self-serve access across Europe.
Organic visibility in ChatGPT is open to every merchant with a clean feed today. The ad auction is now open, self-serve, across the US, UK and 31 European markets, and it's still cheap. A brand that's invisible in ChatGPT's organic answers won't suddenly look credible in a sponsored card.
Lengow channel analysis · How to get ready for ChatGPT Ads in Europe · Is ChatGPT the new Google Shopping?
FAQ
Frequently asked questions
Can’t find what you’re looking for?
Two ways. Organically: ChatGPT Shopping shows product results with images, prices, ratings and merchant links, ranked on availability, price, quality and whether you're the maker or primary seller, powered by a product feed you push to OpenAI. And through ads: a single sponsored card below the answer, clearly labelled, which never changes what ChatGPT itself says.
Any merchant can apply to OpenAI for direct product feed access, and Shopify merchants are integrated automatically through Shopify Catalog. Advertising additionally requires Ads Manager access in a live market: the US (self-serve), Canada, or the UK (via OpenAI representatives).
The organic layer is free: there is no charge for the product feed or for appearing in shopping results. Ads run on CPC (Clicks objective) or CPM (Reach objective) in a relevance-weighted second-price auction, with maximum bids set per ad group; OpenAI recommends starting around $3-5 maximum CPC. Early pricing sits below Google and Meta, which won't last as competition arrives.
Only logged-in adults on the Free and Go plans. Plus, Pro, Business, Enterprise and Edu accounts never see ads, and OpenAI excludes users it knows or predicts are under 18. Organic shopping results, by contrast, appear for any user having a shopping conversation.
A structured file (CSV, TSV, XML or JSON) pushed to a secure OpenAI endpoint and validated before going live: identifiers and GTINs, factual titles and descriptions, accurate prices and availability, variants with stable URLs, reviews where available, seller name and published return, privacy and terms policies. Updates are accepted as often as every 15 minutes, and the enable_search and enable_checkout flags control eligibility per product.
Lengow has supported the OpenAI Product Feed since its launch: catalogue mapping to the specification, enriched titles and attributes for natural-language matching, enable_search and enable_checkout flags managed per segment, automated updates on the channel's cadence, and validation before going live. The same optimised feed then powers feed-based ad campaigns in Ads Manager as markets open, one workflow for the organic and paid layers.
Take control of your advertising performance
Join thousands of brands and retailers using Lengow to structure, optimise and distribute their product feeds across every advertising channel, with data-driven insights and full control.