Klarna announced it was joining Google's Universal Commerce Protocol (UCP) yesterday, Feb 2nd.
I dont think it's a straightforward infrastructure move or simply just payment processing. This is bigger. It's about making Buy Now Pay Later (BNPL) a product attribute in agentic commerce, not a checkout option.
The Protocol Layer That Matters
Google's UCP, co-developed with Shopify, Target, Walmart, and others, establishes a common language for AI agents to interact with commerce systems across the entire shopping journey. The technical architecture separates product discovery, checkout flows, and payment handling into modular, extensible layers.
But here's what makes Klarna's move strategically distinct: UCP enables AI agents to surface payment options during product discovery, not just at checkout. When an AI agent is evaluating products, it can now present "available with Klarna" alongside price, shipping, and availability, all before a user ever decides to buy.
That's not a payment feature. That's demand generation through protocol adoption.
BNPL as Product Metadata
Think about how AI agents will shop. When ChatGPT or Gemini searches for a product, it wont just look at the sticker price, it will evaluate the total cost of ownership, availability, delivery time, and inevitably, how you can pay for it.
Klarna already launched its Agentic Product Protocol in December 2025, giving AI systems access to over 100 million products and 400 million real-time prices across 12 markets. That protocol makes products discoverable. UCP makes the payment terms part of that discovery layer.
When an agent recommends a £800 laptop, UCP integration means Klarna's "real-time decisioning engines" can tell that agent and the shopper, that the purchase is actually "£67/month with Klarna" before the user has even clicked through to a product page.
This is Klarna inserting itself into the "can I afford this?" moment that happens during AI product evaluation, not after.
The "Affordability Metadata" Problem
Here's the technical challenge that makes this strategically important: LLMs don't inherently understand payment flexibility.
A £600 purchase is just "£600" to an AI agent unless someone has structured the payment optionality into machine-readable metadata. Without UCP standardization, every BNPL provider would need custom integrations with every AI platform, every merchant backend, and every product catalog.
UCP's payment handler design solves this through negotiation: merchants declare which payment handlers they support, AI agents query those options, and the protocol mediates what's available for a specific transaction.
Klarna's UCP adoption means their flexible payment options become natively parseable by AI agents, part of the standard product profile that every UCP-compliant agent can read and present to users.
If Affirm, Afterpay, or other BNPL competitors don't adopt UCP (or whatever protocol dominates), their payment options remain invisible during AI-driven product discovery. They're relegated to traditional checkout flows that happen after the AI has already narrowed the options.
First Mover Advantage Through Infrastructure
Klarna isn't the only BNPL provider preparing for agentic commerce. Affirm adopted Google's Agent Payments Protocol (AP2) in October 2025, and its CEO Max Levchin has been vocal about AI agents "weeding out products with prohibitively high interest rates."
But there's a crucial difference between AP2 and UCP: AP2 focuses specifically on payment execution; the secure, agent-led completion of transactions. UCP covers the full commerce journey, from discovery through post-purchase support. It's the difference between "how to pay" and "what payment options exist for this product."
Klarna is betting on both. They adopted AP2 months ago and are now layering on UCP, which means they're positioned at both the discovery and transaction layers of agentic commerce ;)
The strategic implication: if Klarna's payment options are protocol-native at the discovery layer, they become the path of least resistance for AI agents. Not because they're better (though they might be), but because they're structurally easier to understand, surface and recommend.
The Merchant Incentive Misalignment
Here's where it gets interesting for retailers: merchants want any BNPL option that converts browsers into buyers. They don't particularly care whether it's Klarna, Affirm, or Afterpay, they care about closing the sale.
But if only Klarna's payment terms are natively available during AI product discovery through UCP, merchants face an asymmetric integration burden:
- Klarna: Already discoverable through protocol-standard metadata
- Other BNPL providers: Require separate integrations, custom data feeds, or are only surfaced at checkout
This creates subtle lock-in through standardisation, not contracts. Merchants don't need to explicitly choose Klarna, the protocol architecture makes Klarna the default because it's easier to expose through UCP than to maintain parallel integrations for competitors.
Shopify's technical breakdown of UCP makes this clear: "merchants set up their data once, and through Shopify's Agentic Storefronts, that data is surfaced everywhere." If that "once" includes Klarna's payment options as UCP-compliant metadata, merchants get AI-agent discoverability for free. Adding competitors means additional work.
What This Actually Means for Agentic Commerce
Let's be concrete about what this may look like in practice.
Scenario without UCP-native BNPL:
- User: "Find me a good laptop under £1000"
- AI Agent: "I found this £899 Dell XPS. It has 16GB RAM, 512GB SSD..."
- User: "Can I pay in installments?"
- AI Agent: "Let me check checkout options..." [redirects to merchant site]
Scenario with Klarna's UCP integration:
- User: "Find me a good laptop under £1000"
- AI Agent: "I found this £899 Dell XPS (or £75/month with Klarna). It has 16GB RAM, 512GB SSD..."
- User: Makes decision with payment optionality already factored in
The difference is that in the second scenario, BNPL isn't a checkout feature, it's a product attribute. Like free shipping or Prime eligibility, it affects whether the user even considers the purchase worth exploring.
More importantly, Klarna's "real-time decisioning engines" mentioned in their press release could mean instant pre-approval signals during discovery:
"I found this laptop, and you're pre-approved for 12 monthly payments of £75."
That's not just convenience, it's demand generation embedded in product metadata. The AI isn't just showing what's available; it's showing what's affordable for you, in real-time, before you've decided whether to buy. Klarna is positioning for this very situation.
The Strategic Question: Commoditisation or Capture?
This raises a critical strategic question for anyone watching the BNPL space: Is Klarna using UCP to commoditise BNPL (making it standardised, interoperable, and protocol-compliant), or to capture the category by moving first?
The answer is probably both, and that's the genius of the move.
By supporting an open standard, Klarna positions itself as pro-interoperability, merchant-friendly, and aligned with Google's vision of "an agentic commerce future that is open, collaborative and built for everyone to succeed." That's good optics and good politics as regulators eye the BNPL sector.
But by moving first and moving comprehensively (both AP2 for payments and UCP for discovery), Klarna gets structural advantages:
- Default positioning in AI agent recommendation engines
- Lower merchant integration friction compared to non-UCP competitors
- Earlier access to the emerging agentic commerce market as Google rolls out checkout features in AI Mode and Gemini
The companies that will struggle are BNPL providers who sit this out or adopt these protocols slowly. If agentic commerce grows as quickly as McKinsey predicts (£1 trillion in orchestrated revenues by 2030), the infrastructure layer gets locked in early, just like payment rails, search algorithms, and app store guidelines.
By the time competitors realise UCP adoption matters, Klarna may have already become the BNPL option that AI agents "just know about" because it's protocol-native.
Why This Matters More Than You Think
The broader implication extends beyond Klarna or even BNPL: agentic commerce isn't about checkout. It's about product comprehension.
AI agents need to understand not just what a product is, but how it can be acquired: price, shipping, availability, return policy, and increasingly, payment flexibility. The companies that embed their services into that comprehension layer, through protocol adoption, structured data, and API accessibility will win.
Klarna gets this. Their Agentic Product Protocol makes products discoverable. UCP makes payment terms part of the product definition. Together, they position Klarna as infrastructure, not just a payment option.
As Sundar Pichai noted when launching UCP: "The industry needs a protocol that works at global scale and takes into account the nuances of commerce journeys."
For Klarna, those "nuances" include making sure that when an AI agent evaluates whether a product is affordable, Klarna's payment options are part of the answer, not an afterthought at checkout, but a core attribute evaluated during discovery.
That's not a payment integration. That's becoming part of how AI agents think about products.
And that's worth a lot more than transaction fees.