Marketing Personalization · Sales Automation
Magalu wanted to make product discovery, comparison, purchase, and post-sale support faster and more intuitive, replacing complex retail navigation flows with a conversational shopping experience that could still operate at marketplace scale.
Reported impact
Magalu's conversational shopping assistant outperformed the retailer's app on conversion, attracted millions of visitors in its first six months, and gave the company a scalable path to AI-driven retail personalization inside a channel customers already use.
Magalu built Lu's WhatsApp as an end-to-end AI commerce experience that turns shopping into a conversation inside WhatsApp. The interface is powered by a large-scale language architecture that combines proprietary and external models in a coordinated multi-agent system. Trained on data from millions of active users, 300,000 sellers, and more than 37 million listings, the implementation uses more than 20 orchestrated AI agents to interpret search intent, ask clarifying questions, recommend products, support customers, and move users through transactions. Customers can interact with Lu through text, voice, or images, receive curated product options, pay within WhatsApp, and track orders without leaving the conversation. The story does not name a distinct storage layer, but it makes clear the platform relies on Magalu's marketplace data corpus and operational metrics to iterate quickly, manage quality, and scale toward tens of millions of users.
Customer prompts arrive through WhatsApp as text, voice, or images, Lu's WhatsApp passes them into a language architecture and multi-agent system, and the agents use Magalu's marketplace data to search listings, recommend products, support purchases, and return order updates in the same conversation.
"3x higher conversion than the app"
"At its core, Lu's WhatsApp runs on a large-scale language architecture that integrates proprietary and external models into a coordinated multi-agent system."
"Customers can use text, voice, or images to search for products, receive personalized recommendations, complete purchases and pay directly within WhatsApp, and track orders."
"With the launch of the world's first end-to-end AI commerce experience inside WhatsApp"
"the system operates through more than 20 orchestrated AI agents"
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
In-house Sample Jun 18, 2026 Target As shoppers increasingly begin discovery and purchase journeys inside conversational AI tools, Target needed to make its assortment, account identity, loyalty benefits, and checkout available directly in those… AI-driven traffic to Target leapt 2,000% in the first quarter compared to last year. ChatGPT · Copilot Checkout · Gemini app 
Google Cloud Sample Archived · Feb 7, 2026 Verkkokauppa.com We've been a long-time user of Google Cloud, so it made sense to build on our existing data foundation in BigQuery. Google Cloud is a leader in AI innovation, and if anyone understands search, it's Google, so we were… 4% increase in products added to cart BigQuery · Vertex AI · Vertex AI Search 
Bain Sample Archived · Dec 28, 2025 Káyana Metro Pacific Investments Corporation, Manila Electric Company, and PLDT wanted a compliant way to unify fragmented cross-industry datasets so they could generate actionable insights, automation, and advanced analytics… 200 customer micro-segments created AI-powered online acquisition campaign · Káyana · machine-learning-based lead acquisition engine 