Strabo Proves Formal AI Agent Protocols Can Beat Google's Commerce Standard
Researchers show that formal, declarative agent protocols can interoperate with Google's live UCP checkout infrastructure, offering a rigorous alternative to ad-hoc agentic design.
- Strabo shows that formally-specified AI agents built with Langshaw and Peach can interoperate with Google's live UCP checkout infrastructure.
- Google's Universal Commerce Protocol (UCP) is an open standard for agentic commerce across Search, Gemini, and other surfaces, co-developed with Shopify, Walmart, Target, and 20+ partners.
- Langshaw is a declarative protocol language that lets you specify what an agent interaction must achieve, with machine-checkable safety and liveness guarantees, instead of just sequencing API calls.
- The key result: Peach-based agents completed real UCP checkout flows against Google's own implementation, proving formal specs can match production behavior.
- The approach enables incremental adoption: formally-specified agents can be introduced alongside conventional UCP agents without requiring a full ecosystem switch.
- The paper only covers UCP's checkout flow, not the full protocol; no code is publicly released, and performance benchmarking is absent.
As AI agents start handling real commercial transactions, the question of how they should coordinate with each other is no longer theoretical. A new paper called Strabo from researchers at NC State and Lancaster University makes a pointed argument: the academic field of multiagent systems has spent years building formal, provably-correct tools for exactly this problem, and the industry is largely ignoring them. Their proof of concept targets one of the most concrete industry efforts in this space right now: Google's Universal Commerce Protocol.
The Protocol Wars Are Already Happening
Google's Universal Commerce Protocol (UCP) is an open standard designed to let AI agents handle the full shopping journey, from product discovery through checkout, across surfaces like AI Mode in Google Search and Gemini. It is designed for the future of commerce, empowering businesses to turn AI interactions into instant sales. UCP is built to work across verticals and is compatible with existing industry protocols like Agent2Agent (A2A), Agent Payments Protocol (AP2), and the Model Context Protocol (MCP), and was co-developed with industry leaders including Shopify, Etsy, Wayfair, Target, and Walmart, with endorsement from over 20 others.
The core problem UCP solves is fragmentation. Before UCP, if you wanted to buy something through an AI assistant, that assistant needed a custom connection to each individual store. UCP removes that friction entirely, working similarly to how HTTPS standardized secure web communication. By establishing a common language and functional primitives, UCP enables seamless commerce journeys between consumer surfaces, businesses, and payment providers.
What the Research Community Brings to the Table
The Strabo paper's argument is that UCP, while practical and well-engineered, was built without the benefit of decades of formal multiagent systems research. That research field, known in academic circles as EMAS (Engineering Multiagent Systems), has developed rigorous, mathematically grounded tools for specifying how agents should interact. The Strabo team asks: can those tools be applied to something as real-world as UCP, and can the resulting agents actually talk to Google's existing implementation?
The two key tools they use are Langshaw and
This story is for Pro members
You've reached the end of the free preview. Upgrade to AlphaSignal Pro to read the full article - and everything else behind the paywall.
Comments
