Skip to main content

End of shopping chaos? Adyen introduces a 'universal translator' for the era of AI agents

Ilustrační obrázek
The consumer's purchasing journey has just invisibly changed. While previously we searched for products using keywords and then manually went through shopping carts, we are entering the era of agentic commerce. In this new reality, people no longer have to do everything manually – autonomous AI agents decide, negotiate, and purchase on their behalf. For companies, however, this presents a technological catch: how to sell goods to someone who is not a human, but software with its own protocols? FinTech leader Adyen has just answered with its new platform, Adyen Agentic.

What is "Agentic Commerce" and why should you care?

The term agentic commerce refers to a phase of digital commerce where the purchasing process is not driven solely by human-interface interaction, but primarily by AI agent-merchant interaction. Imagine it this way: instead of you yourself searching in a browser for "best ergonomic chair under 5000 CZK", you tell your personal AI assistant (e.g., ChatGPT or a future version of Google Gemini): "Find me the best chair, compare reviews, and order it using my payment card."

This process requires the merchant to be visible not only to people on websites but also to be "readable" by algorithms. As Adyen states, we can move between two states: conversational commerce, where AI merely assists a human with selection, and a fully autonomous era, where agents handle transactions themselves at "machine speed."

Adyen Agentic: Solving the Fragmentation Problem

For large retailers, the biggest challenge is that every AI agent (whether it's OpenAI, Google, or specialized startups like New Gen) uses different protocols, different product data formats, and various requirements for creating a shopping cart. Without a unified standard, every company would have to create new integrations for each new AI assistant that appears on the market.

Adyen Agentic functions as a "universal translator." Companies perform only one integration into the Adyen system and can then immediately communicate with the entire ecosystem of AI agents. Among the first partners supporting this system are financial giants like Visa, Mastercard, and American Express, as well as technology leaders like Salesforce.

This approach eliminates the need for "bets" on a specific platform. A seller doesn't have to wait for one particular AI model to win; thanks to Adyen, they can be present wherever agents operate.

Key Acquisition: Orb and Usage-Based Billing

Part of this strategy is also the recent acquisition of Orb for 335 million USD. Why is this important? Traditional e-commerce operates on the principle of one-time payments for a product. However, AI agents often work in usage-based pricing models. For example, when purchasing software services or cloud storage, an agent can negotiate the price based on the actual consumed amount of data or tokens.

By combining Adyen's payment infrastructure and Orb's capabilities for complex billing, Adyen can automate revenue decision-making in real-time. This is a fundamental shift for companies moving away from fixed subscriptions towards dynamic payment for actual value.

Practical Impact: What Does This Mean for the Czech Market and the EU?

Although Adyen Agentic is currently primarily focused on large enterprises in the USA, its impact on the European and Czech markets will be inevitable. For Czech e-shops and technology companies, this has three main aspects:

  1. Availability and Integration: Adyen is a global player with a very strong presence in Europe. Once the solution expands, Czech retail (e.g., large chains or specialized e-shops) will have to adapt their systems to be "agent-ready."
  2. Regulation and Security: In the EU, the AI Act and strict GDPR regulations play a primary role. Autonomous agents that purchase on behalf of people handle sensitive payment data and personal preferences. Security protocols must be extremely robust to prevent risks of unauthorized transactions carried out by AI.Competition for Traditional Websites: Startups like New Gen are already building "AI storefronts" – dynamic versions of websites that are optimized for bots, not for people. This means that a Czech e-shop with only a static catalog might completely disappear in the eyes of AI assistants.

    Comparison: Where Are We Now?

    If we compare the current state with technologies like GPT-4 or Claude 3.5, we see a clear trend. While these models are excellent at generating text and analyzing data, for real commerce they need "hands" – i.e., payment interfaces. Adyen strives to be these hands for the entire world of AI agents.

    Feature Traditional E-commerce Agentic Commerce (Adyen)
    Main User Human (clicking, browsing) AI Agent (algorithm, instructions)
    Search Keywords in Google/Seznam Intent and Context
    Integration API for web browsers Universal protocol translator

    Conclusion

    The future of commerce is not about better website design for people, but about the ability of machines to communicate with each other. Adyen Agentic strives to build an infrastructure that will enable this transition without chaos. For companies, it is an opportunity for massive automation, but also a challenge to remain visible in a digital world that will no longer be controlled solely by the human eye.

    Is it safe to let an AI agent shop for me?

    Security depends on the implementation of protocols. New systems like Adyen Agentic work with certified payment channels (Visa/Mastercard) that must meet strict security standards and regulations like GDPR in the EU to ensure control over what amounts and for what an agent can spend.

    Does my company have to be large to use these technologies?

    Currently, Adyen Agentic is primarily targeted at the enterprise segment (large corporations). However, with the development of the ecosystem and tools like Orb, these possibilities will likely gradually open up to smaller companies through SaaS platforms.

    How do I know if my e-shop is ready for AI agents?

    The foundation is structured data. If your product data is not clearly defined in machine-readable formats (e.g., via JSON-LD or specific APIs), it will be very difficult for an AI agent to find your product and correctly interpret its parameters.

X

Don't miss out!

Subscribe for the latest news and updates.