Criteo Debuts AI Commerce Recommender

Criteo's new "Agentic Commerce Recommendation Service" aims to power AI-driven shopping assistants with more relevant product suggestions. This service connects AI chatbots and assistants directly to a retailer's live inventory, using real-world shopping behavior rather than just static product descriptions to inform its recommendations. The company's internal tests from January 2026 showed up to a 60% improvement in recommendation relevancy compared to systems that rely only on product descriptions. This performance is backed by Criteo's large-scale data, which includes 720 million daily active shoppers and insights from $1 trillion in annual e-commerce transactions. The system filters products based on popularity, availability, and user intent to provide a curated list. This launch positions Criteo in the growing AI in retail market, which is projected to expand from approximately $14.24 billion in 2025 to $96.13 billion by 2030. The move pits them against established tech giants like Google, Amazon, and Microsoft, as well as other ad tech firms like Taboola and Quantcast, all competing to provide the AI-powered personalization layer for e-commerce. The service is delivered through Criteo's Model Context Protocol (MCP) and has been in testing with a major large language model (LLM) platform since 2025. Criteo's CEO, Michael Komasinski, stated that the competitive edge in "agentic commerce" will come from access to high-quality, large-scale commerce data. As AI personalization becomes more sophisticated, retailers face the challenge of balancing customized experiences with consumer data privacy. A recent McKinsey report noted that while 71% of consumers expect personalized interactions, they are also increasingly concerned about how their data is used. Striking this balance is crucial for maintaining customer trust while leveraging powerful AI tools.