Klarna as AI Commerce Middleware

Klarna is trying to become the product feed and demand router that sits before checkout, not just the lender that shows up at checkout. Agentic Product Protocol turns Klarna’s catalog, pricing, and availability data into infrastructure that AI assistants can query directly, so a merchant using Klarna can be discovered inside agent interfaces, then monetized through sponsored placement, affiliate economics, and eventually payments when the purchase happens.1 2 3 4

• Klarna already makes money from merchant ads and shopping traffic, not only BNPL. In 2024, BNPL merchant fees were 57 percent of revenue, while in app merchant ads were about $180M and up 15x since 2020. Its filing also says advertising revenue comes from search solutions and affiliate and brand ads on its network.2 4
• The protocol makes Klarna useful in a new place in the commerce stack. Instead of waiting for a shopper to reach a merchant checkout page, Klarna can now supply normalized product data, live prices, images, brands, and listings to AI systems earlier in the shopping flow, across 100M plus products and 12 markets.1 3
• This looks more like universal checkout and commerce middleware than a pure credit product. Bolt’s core pitch was that a shared shopper and merchant layer could sit across many storefronts and sell more software on top. Klarna is applying a similar network logic one step higher, at discovery and recommendation inside AI channels.5 6 7

The next step is for Klarna to bundle product distribution, ad spend, and payment acceptance into one merchant stack. If AI agents become a real shopping surface, merchants will want the same company to supply the catalog feed, buy visibility, track conversion, and close the transaction. That creates a larger, more defensible business than BNPL alone.1 2 4 7