Edgebound Lab
Where commerce meets artificial intelligence.
We build prototypes, experiment with emerging protocols and publish what we learn. Agentic Commerce, AI applied to search and pricing, personalization models: if it is redefining how things are bought and sold, we are researching it.
It is not innovation theater. It is running code, real data and documented results.
What we are researching
AI agents are learning to buy. ACP (OpenAI + Stripe) and UCP (Google + Shopify) are the protocols that define how an agent discovers products, negotiates prices, executes payments and handles returns, with no human in the loop.
We are implementing both protocols, documenting the technical differences and exploring how they operate in the context of enterprise commerce for LATAM.
We embed artificial intelligence into every layer of the commerce stack: semantic search with retrieval-augmented generation (RAG), dynamic pricing with demand models, real-time content and product personalization, and customer support with agents that know the full catalog.
We do not experiment with AI in the abstract: we apply it to real metrics: conversion, AOV, retention.
What we are building
Agent that runs the full checkout flow via ACP: product discovery, cart and payment with Stripe. Built on Shopify Plus.
Discovery engine that consumes UCP capability profiles to compare products cross-merchant in real time. Agent-agnostic: works with GPT, Claude and Gemini.
Semantic search over commerce catalogs with retrieval-augmented generation. Understands intent, not just keywords. +43% conversion vs. traditional search.
Dynamic pricing model that adjusts prices in real time based on demand, inventory, competition and elasticity. Integrated with Commercetools and Shopify.
Interactive digital maturity diagnosis for eCommerce. 3 questions, instant result with personalized recommendations. No email required.
Controlled environment where AI agents interact with a test catalog via ACP and UCP simultaneously. To validate interoperability between protocols.
Agentic Commerce: ACP vs. UCP
This is the comparison from the perspective of whoever implements them.
Enterprise merchants will need both protocols. ACP to capture ChatGPT traffic today; UCP to be ready for tomorrow's multi-agent ecosystem. We are building the reference implementations for both.
Prototype · Live
Three questions. A data-driven diagnosis. We tell you where to start.
—
Prototypes, not PowerPoints
We do not experiment from scratch. Every prototype is built on two decades of implementations for Costco, IBM, Johnson & Johnson and enterprise clients across LATAM.
We are not locked to a vendor. We implement ACP and UCP on Shopify, Commercetools and custom stacks. If the protocol exists, we integrate it.
Experimentation does not compromise security. Every prototype operates under our certified security management system.
The lab's vocabulary
If you are evaluating agentic commerce for your operation, or want to understand how AI can impact your commerce metrics, let's start with a technical conversation.
contacto@edgebound.com · Mexico City, Mexico