Unilever Revolutionizes E-Commerce with Multi-Agent AI Customer Engagement

Unilever, a global consumer goods company, faced the challenge of complicated website navigation and low conversion rates due to customers struggling to find products and information efficiently. In partnership with Capgemini and Microsoft, Unilever implemented a multi-agent AI system built on Microsoft Azure, orchestrating a suite of specialized AI agents to deliver seamless, contextual, and multi-language customer experiences on their e-commerce platform. The multi-agent system leverages Azure AI, Azure OpenAI (GPT-4), Azure AI Search, Cosmos DB, and PromptFlow to power natural language interactions, personalized product recommendations, and adaptive global language support. An orchestrator agent intelligently routes queries to the right agent, while retrieval-augmented generation (RAG) enables the system to combine internal product and recipe data with external generative models for accurate, real-time responses. Key safety guardrails were implemented, including robust content moderation and prompt restrictions, ensuring secure interactions across use cases. The solution not only enhanced discovery and recommendation features but also improved order fulfillment by integrating with back-end shopping cart workflows. Notable business outcomes include faster and more accurate query responses, increased user engagement, improved conversion rates, and a frictionless global user experience, setting a new e-commerce standard for digital customer care.