Agentic commerce currently runs on four protocols. OpenAI's ACP, Google's UCP, AP2, and Meta's checkout system each impose different data formats, different authentication requirements, different checkout logic. A merchant who wants AI agents to discover and purchase their products has to pick which protocols to support — and there's no consensus on which will win.

On June 16, Adyen announced a product designed to sidestep that choice. Adyen Agentic is a three-layer API suite: catalog data distribution, checkout orchestration, and payment execution. One integration, compatible with all four protocols. The company positioned it as "the universal translator for the next era of commerce."

What made the launch event unusual was who showed up to endorse it. American Express, Mastercard, and Visa — all three major card networks — were present and on record backing the same neutral infrastructure layer. Salesforce was there too. Getting direct competitors to jointly endorse the same platform doesn't happen often. It likely reflects something specific about the current moment: none of them want to own the protocol layer themselves, but all of them have strong reasons to ensure agentic payments work reliably across the board.

The Announcement: One Integration to Rule Them All

On June 16, 2026, Adyen announced Adyen Agentic in New York City, pitching it as a solution to one of agentic commerce's most pressing structural problems. A merchant integrating with OpenAI's Agentic Commerce Protocol (ACP) gets access to one pool of AI shoppers. Google's Universal Commerce Protocol (UCP) is another. Meta's AI checkout flow is a third. AP2 a fourth. Each demands different product data formats, different authentication handshakes, different checkout orchestration logic. Merchants can't practically rebuild their commerce stacks four times over.

The premise of Adyen Agentic is elegant in its ambition: integrate once with Adyen — a payments infrastructure company already embedded in the checkout flows of thousands of enterprise merchants — and your stack becomes immediately compatible with every major agentic shopping protocol simultaneously.

Karan Katyal, Adyen's Global Head of Agentic Commerce, put it directly: "We believe the future of agentic commerce should be open, so we intentionally designed Adyen Agentic to help retailers integrate once and participate across evolving platforms, protocols, and experiences."

The Architecture: Three Layers, One Integration

Adyen Agentic is structured around three discrete API layers that map to the full journey of an AI-mediated purchase — from the moment an agent asks a question to the moment a transaction clears.

Agentic Feed — the data layer. Structures and distributes real-time catalog, pricing, and inventory information into conversational commerce environments. When a shopping agent asks "what trail running shoes do you have in size 10, available for next-day delivery to Denver?", the answer has to come from somewhere, formatted in a way the AI can parse and act on. Agentic Feed is the infrastructure that makes merchant data machine-readable, current, and trustworthy for agent consumption.

Agentic Cart — the orchestration layer. Connects AI agent purchase intent to the merchant's existing checkout, tax, fulfillment, and order management systems. This is where the operational complexity lives: reconciling agent requests with real-world inventory constraints, tax jurisdictions, shipping rules, and fulfillment capacity — systems that were never designed to receive instructions from autonomous software.

Agentic Payments — the trust layer. Manages token portability, authentication, fraud detection, risk modeling, and merchant-of-record responsibilities for agent-led transactions. This is where Adyen's core payments heritage is most visible: ensuring that when an AI agent executes a purchase on a human's behalf, the right delegated tokens, authentication signals, and risk assessments fire correctly — with no human in the loop at the moment of transaction.

The Ecosystem Play: When Your Competitors Both Sign On

What makes Adyen Agentic architecturally significant isn't just the three-layer design. It's the list of names who endorsed it at launch.

American Express. Mastercard. Visa. Salesforce.

All three major card networks simultaneously backed the same protocol-neutral infrastructure layer. Sherri Haymond, EVP at Mastercard, emphasized that "trust, security and interoperability need to be built in from the start." Rubail Birwadker, SVP at Visa, echoed: "embedding trust, security and global acceptance into every transaction."

The protocol compatibility list reinforces the neutrality: Adyen Agentic explicitly supports Google's Universal Commerce Protocol (UCP), OpenAI's Agentic Commerce Protocol (ACP), Agent Payments Protocol (AP2), and Meta's AI checkout — the four leading standards in agentic commerce, representing three competing technology empires.

This is not a bet on one protocol winning. It is, explicitly, a bet that no single protocol will dominate, and that merchant value lies precisely in not having to choose.

Real Merchants Already Live

Adyen Agentic is not a research preview or a pilot. Early enterprise merchants are already operating on the platform: ESW (cross-border commerce specialist), Scheels (sporting goods retailer), Sézane (French fashion brand), and SharkNinja (consumer appliances and home goods).

Nicolas Benoist, CTO at Sézane, captured the strategic logic: "build on proven, battle-tested foundations to deploy quickly and confidently." For established retailers with functioning commerce stacks, the appeal of adopting agentic compatibility through an existing, trusted payments partner — rather than building proprietary agent integrations — is considerable. The service launched in limited form to U.S. enterprise merchants, with global expansion planned.

The Orb Acquisition: Closing the Loop

The Adyen Agentic announcement didn't land in isolation. Days before, Adyen agreed to acquire Orb — a billing infrastructure platform specializing in usage-based, consumption-based pricing — for $335 million. Co-CEO Ingo Uytdehaage framed the deal as "closing the loop between what merchants charge and how those charges perform."

The combination hints at where Adyen sees agentic commerce heading: toward metered, incremental transactions where AI agents purchase data, services, and physical goods in small units, at high frequency, across many channels simultaneously. Traditional e-commerce billing — a discrete purchase event with a clear SKU and price — doesn't cleanly model a world where an agent might assemble a cart across six merchants over several hours before triggering a single optimized checkout.

Together, Adyen Agentic and the Orb acquisition suggest a coherent bet: position as the full-stack infrastructure layer for a world where AI agents buy things continuously, granularly, and autonomously.

The Infrastructure Race Heating Up Around It

Adyen Agentic is the week's headline, but it landed amid a broader sprint to lay down agentic commerce rails from every direction.

Mastercard's Agent Pay for Machines (AP4M), launched June 10, introduced something architecturally distinct: blockchain-anchored agent permissions recorded on Polygon, Solana, and Base — enabling machine-to-machine micropayments at fractions of a cent. With 30+ launch partners including Adyen, Coinbase, and Cloudflare, AP4M targets the emerging autonomous economy where AI agents pay each other for data and services in real time. Mastercard CPO Jorn Lambert offered a notably tempered but directionally bullish read: "Do I think it'll be a meaningful new addressable market for us over the next five years? I think so."

Paymentology CTO Tim Joslyn went further, predicting that within three years, AI agents could autonomously manage procurement, treasury, and operational spending with minimal human oversight.

Shopify's Spring '26 Edition, released June 17 with 150+ updates, declared agentic commerce as production-ready default infrastructure — making merchants discoverable by AI agents by default, not by opt-in configuration.

At Shoptalk Europe 2026, the traction numbers were striking: Zalando's AI shopping assistant grew from 6 million users in 2025 to 10 million in Q1 2026 alone. The platform's AI-generated site content went from near-zero to 90% of total content within a year. These are not marginal pilot metrics.

The Challenge Beneath the Momentum

Despite the infrastructure sprint, the conversion reality still points to early innings. Just 2.2% of AI agent interactions currently reach the cart or payment stage — the vast majority terminate at research and comparison browsing. Forrester's mid-2026 assessment of agentic commerce identifies an "AI execution gap" where adoption is outpacing measurable business impact.

Consumer trust remains the structural constraint. Buyers are increasingly comfortable delegating research to AI agents, but significantly more reluctant to extend autonomous payment authority. The question of "who is liable if the agent buys the wrong thing?" hasn't been cleanly resolved by any protocol — including Adyen Agentic.

78% of financial institutions anticipate fraud spikes from AI agents exhibiting novel transaction patterns that confound traditional behavioral detection systems. Adyen's trust layer addresses part of this at the transaction level. But the deeper trust gap — human willingness to pre-authorize purchasing decisions to software — is behavioral and regulatory, not just infrastructural.

What the Next Chapter Looks Like

The "universal translator" proposition holds as long as the protocol landscape remains fragmented. If one protocol — UCP, ACP, or an emerging challenger — achieves genuinely dominant adoption, the integrate-once value proposition weakens accordingly. Adyen's bet is that protocol complexity stays high long enough for the infrastructure abstraction layer to become indispensable — and that, once merchants integrate, switching costs do the rest.

That logic has strong historical precedent in payments. The global payments industry runs on a fragmented web of card networks, bank rails, local schemes, and digital wallets — and the companies that built protocol-neutral abstraction layers above that complexity have generated enormous durable value.

Agentic commerce may follow the same infrastructure pattern. The question isn't whether AI agents will increasingly mediate commerce — the Zalando numbers, the 15x growth in AI-driven orders over 2025, and the Shopify default settings all suggest that ship has sailed. The question is who builds the roads everyone else drives on.

Last Monday, Adyen made a very explicit bid for that role.

Sources