BOSTON – Cimulate AI announced the introduction of a proprietary platform that redesigns product discovery via the first commerce-optimized large language model (LLM) built for agentic commerce.
Cimulate CommerceGPT is the first AI-native platform built for agentic commerce, a phrase that refers to the use of AI-powered agents to autonomously handle shopping tasks, from product discovery to purchase and even post-purchase activities like returns.
The company says Cimulate CommerceGPT facilitates the paradigm shift now underway in e-commerce as consumers cede direct control of some or all steps in the shopping experience, delegating tasks to intelligent software agents.
“Our proprietary approach — distillation via simulation — generates synthetic, high-fidelity shopping behavior based on LLMs and your product catalog,” says John Andrews, Cimulate’s co-founder and chief executive officer. “It simulates customer behavior. It closes the gap. It gives your search, recs, and conversational co-pilots the context they’ve always lacked. This new engine doesn’t just improve product discovery, it redefines it.”
Built to understand shopper intent and drive conversion
Unlike Bing, Google and other traditional search engines, LLMs are built to understand language, infer intent and reason about goals, the company says. “They don’t just match keywords – they interpret meaning. They don’t just recall behavior – they respond to nuance.”
But while general-purpose LLMs are too broad and too expensive to deliver contextual commerce. Cimulate CommerceGPT is optimized to get around these problems; it’s built to understand shopper intent and drive conversion, the company says. “We start by training a lightweight retrieval model on each customer’s product catalog and behavioral data, then enrich it with synthetic shopping-relevant knowledge from general-purpose LLMs. This enables us to simulate millions of realistic shopping sessions, test outcomes at scale and learn what actually converts.”
Cimulate is also introducing MCP Server, a tool to help retailers and brands optimize agent-to-agent commerce.
“While agentic commerce has quickly become a popular term associated with agents supporting the human prompt, a new future is quickly arriving where agents speak to each other following a human prompt. And retailers need to prepare to speak fluent agent – which is both a new opportunity and significant challenge,” the company says. “Our approach helps retailers speak to answer engines like ChatGPT, Perplexity, and Claude in the unique, natural language ways which are completely different than the marketing tactics used to appeal to two or three words in a Google search.”
