Klarna has unveiled a new open technical standard, known as the Agentic Product Protocol, aimed at facilitating artificial intelligence systems in accessing and interpreting vast quantities of online product information.

This initiative is intended to enable AI-driven tools to identify, compare, and recommend consumer goods using structured, real-time data supplied directly by merchants.

The protocol supplies AI agents with access to a continuously updated feed covering over 100 million products and around 400 million price points across 12 markets. Klarna asserts that the data is standardized to include pricing and availability, allowing AI systems to work with live commercial information instead of static or scraped listings.

According to representatives from Klarna, consistent product understanding is crucial for AI systems before they can facilitate transactions. They describe the protocol as a shared framework enabling merchants, platforms, and AI developers to exchange product data in a uniform format, thereby reducing repeated integrations or bespoke listings.

Structure of the Protocol

The Agentic Product Protocol is delivered via a hosted application programming interface (API) managed by Klarna. Merchants can connect their existing product feeds, including formats from major e-commerce and advertising platforms, without restructuring their data.

Upon connection, products become accessible to AI assistants and comparison tools supporting the protocol, without requiring paid placement or intermediary services.

Klarna has announced that both the protocol specification and API access are immediately available to developers, AI platforms, and merchants. As a digital bank and payments provider, Klarna offers installment and deferred payment services across online and in-store channels, including through mobile wallets. The company reports having over 114 million active users globally and processing millions of transactions daily.