On May 20, 2026, Klarna launched a shopping search application inside ChatGPT, according to an official press release published on Businesswire. The app lets a user describe the product they are looking for in natural language, then displays visual results with prices, availability and offers from multiple merchants, all inside a single conversation. Product discovery just moved into the most-used AI interface on the market, and the question for online retailers is no longer "will this happen" but "is my catalog part of it".
What this means in practice: ChatGPT becomes a product discovery surface. Your listings are no longer only indexed by Google, they can now be served to a shopper at the moment of purchase intent, inside a conversation. If your catalog data is incomplete or poorly structured, you are simply absent from the result.
What Klarna actually launched
The app is built on Klarna's Product Search MCP server, which uses the Model Context Protocol to give ChatGPT real-time access to catalogs, prices and stock. According to Klarna's release, this server connects ChatGPT to more than 100 million products spread across 400 million listings in 13 markets.
The flow is simple: the shopper describes their need, Klarna surfaces relevant products with up-to-date prices, then redirects to the merchant's website to complete the purchase. Klarna reports 119 million active users and 3.4 million transactions per day, which gives a sense of the demand volume that can be funneled into this new entry point.
"A consumer who last week would have spent 20 minutes comparing tabs now gets a real answer in one conversation," said David Sykes, Klarna's chief commercial officer, in Digital Commerce 360's coverage. That is exactly the shift to understand: the comparison stage is leaving traditional search engines.
Why your product visibility is on the line
Here is the point retailers should focus on: according to Klarna, merchants appear in organic results ranked by relevance, with an option for clearly labeled sponsored placements. In other words, there is now an organic ranking inside ChatGPT, and it runs on a relevance logic, not just on bidding.
What "relevance" means concretely: the quality of your product data. A precise title, a complete description, structured attributes, accurate prices, up-to-date availability. Retailers who treat their product feed as a technical chore will be at a disadvantage against those who treat it as an editorial asset. The same rule that applies to Google's shopping AI Overviews applies here: AI cites what it understands, not what is merely published.
The context raises the stakes. Still according to Digital Commerce 360, traffic from AI platforms to e-commerce sites grew nearly 700% during the late-2025 holiday season, and those visitors converted 31% better than average. This is not an experimental channel, it is a high-intent one. And the in-app ChatGPT checkout experience shows that the earliest formats do not always convert well, all the more reason to get the discovery stage right, where the decision is actually made.
Are your products ready to be cited by ChatGPT?
Cicéro audits your catalog's visibility on Google, ChatGPT and AI engines. From €250 to €1,800 / month, agency-quality work with software-grade productivity.
What to do now
These three actions have a direct impact on how citable your catalog is inside AI shopping interfaces:
- Audit the quality of your product feed. Descriptive titles, populated attributes (material, size, color, use case), prices and stock synced in real time. An MCP server does not guess, it serves what is explicit. A vague listing is an invisible listing.
- Structure your product pages with
Productschema.ProductandOfferJSON-LD markup remains the most reliable way to make price, availability and reviews machine-readable. It is the foundation of a ChatGPT and AI Overviews visibility strategy. - Write descriptions for humans AND machines. AI ranks by semantic relevance: a description that explains who the product is for, in what situation to use it and what sets it apart will be understood better than a keyword list. Tie this work into your broader GEO strategy for AI search rather than handling it in a silo.
What this news does not cover
A few clarifications to avoid over-reading. The Klarna app is not, at this stage, uniformly rolled out worldwide: coverage depends on the 13 markets cited and on third-party app availability inside ChatGPT. This article does not cover the detailed mechanics of sponsored placements or their pricing, which Klarna has not publicly disclosed. Finally, this is a launch: the traffic volumes actually addressable to any given merchant remain to be measured over the coming weeks. Treat it as a strong signal, not as an already-mature channel.
Our take
Klarna is not just adding a feature: it is moving the purchase-comparison stage into a conversation, and installing an organic ranking inside it. For a retailer, this is the same turning point as the arrival of SEO twenty years ago, but compressed into a few months. The head start belongs to those who clean up their product data now, not to those who wait until they see competitors cited in their place.
Sources
- → Klarna, official press release — May 20, 2026: announcement of the Shopping Search app launch in ChatGPT
- → Digital Commerce 360 — May 20, 2026: MCP server details, David Sykes quote and AI traffic figures
- → PYMNTS — May 2026: analysis of how the app works and real-time pricing
Growth and SEO content strategist, I founded Cicéro to help businesses build lasting organic visibility, on Google and in AI-generated answers alike. Every piece of content we produce is designed to convert, not just to exist.
LinkedIn