.png)
AI Business Mentions First Insight | Retailers Pilot Predictive AI Copilot 'Ellis'
The new retail-focused large language model promises to turn consumer signals into real-time pricing and planning decisions.
AI retail company First Insight has released the beta version of its new AI "growth copilot" Ellis, designed to predict shopping trends months in advance.
The platform is powered by what the company says is the first predictive retail large language model (LLM) developed entirely in-house.
The system is being piloted by several companies as they gear up for the holiday season, ahead of a full public launch in January 2026.
Ellis acts as a conversational interface for the company's decision intelligence platform, enabling retail teams to ask questions in natural language and get predictive, data-backed answers on everything from pricing thresholds to assortment planning.
Unlike general-purpose models trained on web text, Ellis is a domain-specific retail LLM, trained to analyze product and consumer data across categories including apparel, grocery and home goods, and translate consumer feedback into actionable decisions, according to the company.
"Retailers face nuanced challenges that general LLMs, trained on scraped web text, cannot address," First Insight's CMO Viki Zabala told AI Business. "When it comes to delivering measurable retail outcomes, retail LLMs outperform generic ones every single time."
Beta users will test Ellis across five key growth levers: planning, product validation, pricing, demand forecasting and go-to-market optimization.
First Insight stated the tool can compress trend-to-market cycles from nine months to as little as four weeks. For retailers navigating tight budgets and volatile demand, the company pitches Ellis as a turning point where AI moves from insight to immediate action.
Looking ahead, Zabala said Ellis will evolve alongside customer behavior to become a "system of outcome."
"Ellis will evolve to become an orchestrated decision partner that connects insights and actions across systems, making design, planning, pricing and assortment platforms more intelligent and adaptive," she said. "The goal is simple: to help retailers operationalize AI across their ecosystem and move from knowing to doing, faster, smarter and with confidence."
Read on AI Business.
