Vendor transformation research. We made a traditional marketplace agent-ready.
Vendor transformation · Applied research
Summary
We used ReKnew Marketplace to study how an established commerce system could support agent-led shopping without replacing its underlying platform. The work began with a large, flat grocery catalog and a traditional MVC application. It ended with structured product data, headless commerce services, protocol endpoints and tools an agent could use.
Why we built it
Most retailers will not begin agentic commerce with a clean-sheet platform. Their catalog, checkout and order logic already live inside systems built for storefronts and human navigation. We wanted to understand the practical transformation work between that reality and an agent-ready merchant.
What we transformed
We cleaned and enriched more than 2,000 product records, mapped their categories, attributes, prices, inventory and supplier information into a structured commerce model, and rebuilt the customer-facing marketplace around that data. We then exposed the existing catalog, customer, cart, address, payment and order services through a headless REST layer.
Starting system
- Flat catalog
- Traditional commerce logic
ReKnew transformation
- Clean and map the data
- Expose commerce services
- Add discovery and protocol contracts
Agent-ready
- REST · UCP · ACP
- MCP tools
What we tested
On top of the headless services, we built and tested UCP- and ACP-compatible discovery and checkout flows, including the well-known merchant profile used for capability discovery. We also wrapped commerce actions as MCP tools so an agent could search products, work with a cart and move toward an order through a consistent tool interface.
Learning outcomes
The hardest part was not adding an agent interface. It was creating reliable product structure, preserving state across commerce actions and translating existing business services into contracts an agent could safely call. The experiment showed that the underlying platform can remain in place when its data and operations are made explicit, testable and machine-readable.
How this helps our clients
This work became a reference for merchant transformation. It helps us assess where a retailer's current catalog and APIs are usable, where data cleanup or attribute mapping is required, and how to introduce protocol endpoints and MCP tools without rebuilding the entire commerce stack.