What is Klarna’s Agent Protocol Doing for Agentic Commerce?

Klarna has unveiled an open standard designed to make products discoverable by AI agents as AI becomes a popular primary interface for commerce.

The Agentic Product Protocol provides AI systems with access to a live, structured feed containing more than 100 million products and 400 million prices standardised across 12 markets.

The protocol establishes a foundation that allows agents to find, compare and recommend products with live prices and availability across merchants, markets and platforms.

Building the foundation for agentic commerce

Having already established itself as an AI-friendly platform, Klarna expands into agentic commerce. 

"Before agents can buy, they need to know what exists," says David Sykes, Chief Commercial Officer at Klarna.

"Klarna's Agentic Product Protocol defines a common language for how AI systems, merchants and platforms exchange product data: a foundational layer for the next generation of agentic commerce."

The protocol addresses a fundamental challenge in agentic commerce: enabling AI systems to access and understand product information in a standardised format across different merchants and platforms.

Through the hosted API, merchants can connect once to reach any AI agent or platform supporting the protocol, with no reformatting or new listings required.

Supporting multiple feed formats

The system ingests multiple feed formats including Google Merchant, Shopify, Amazon, Facebook Catalog and CSV/JSON feeds.

This flexibility allows merchants to integrate their existing product catalogues without rebuilding their infrastructure or creating separate listings for AI-driven commerce.

Merchants using the protocol can have their products surfaced by AI assistants, making them instantly discoverable and comparable in agent-driven conversations without advertisements, paywalls or intermediaries.

The specification removes the need for merchants to maintain separate integrations with individual AI platforms, reducing technical complexity and expanding reach simultaneously.

Open access for developers

The Agentic Product Protocol specification and API access are now available for developers, AI platforms and merchants.

The open standard approach allows any organisation to implement the protocol without proprietary restrictions or licensing requirements.

This positions Klarna as an infrastructure provider for AI-driven commerce rather than maintaining a closed ecosystem that limits merchant participation.

The protocol focuses specifically on product discovery and data standardisation, establishing the informational foundation that precedes transaction execution in agentic commerce workflows.

Klarna's strategic expansion

Klarna began trading on the New York Stock Exchange in September 2025 under the symbol KLAR, listing 34,311,274 ordinary shares following a period of significant strategic repositioning.

The buy-now-pay-later provider expanded its partnership with JPMorgan Payments in February 2025, integrating its payment methods into the bank's merchant services infrastructure serving 150 million active users globally.

The company recorded first-half 2024 revenue of 13.3bn Swedish crowns (US$1.46bn), representing 27% growth, while US market performance showed 38% revenue expansion in the same period.

Klarna returned to profitability with an adjusted profit of 673 million Swedish crowns (US$61.56m) in the first half of 2024, contrasting with a 456m crown (US$41.7m) loss in the previous year.

Growing field of agentic commerce

The launch comes as major players establish infrastructure for AI-powered transactions.

Visa and Amazon Web Services partnered in December 2025 to bring Visa's Intelligent Commerce platform to AWS Marketplace, aiming to support developers building agentic commerce solutions.

That collaboration introduces tools and open blueprints that enable intelligent agentic workflows and AI-powered digital transactions across sectors including retail, travel and business-to-business operations.

Agentic commerce refers to digital transactions managed by autonomous AI agents that can reason, make decisions and execute transactions without human involvement.

Users can instruct an AI agent to complete specific tasks, such as purchasing products when prices fall below certain thresholds, with the agent carrying out the request autonomously.