La Redoute and Microsoft: Leveraging AI Agents to enhance customer experience

From furniture to ready-to-wear to decoration, for nearly two centuries, La Redoute has made French style its trademark. If the company is now one of the leaders in e-commerce for fashion and home with 36 million monthly visits, it is because it strives to perfect its expertise over the course of evolving trends—and new technologies.

From this dynamic comes multiple innovations, such as a conversational agent for customer relations based on Microsoft technologies, introduced by Jean-Cédric Costa, Chief Information Officer, and Sophie Brief, Head of Customer Service.

A strategic partnership for an ambitious roadmap

At the origin of the project is an ambitious roadmap associated with highly efficient tools, including products developed by Microsoft, La Redoute's partner since 2022. "This partnership began with the launch of our Road to Cloud towards Azure, continuing with the overhaul of our data and AI platform and the implementation of advanced cybersecurity programs," says Jean-Cédric Costa.

Particularly fruitful, this collaboration takes a new turn when generative AI enters the market. Quickly, the group questions the contribution of this technology and how it can best serve its interests. The technical teams, executives, and department heads then meet with one goal: to identify truly adapted use cases. "The idea is not to use technology for the sake of technology, but to identify concrete uses in line with our activity," explains Jean-Cédric Costa.

In this context, customer experience receives the majority preference. The aim? To enhance customer satisfaction by giving them the ability to contact and interact with the company 24/7, using a conversational agent based on Microsoft technologies. "The various projects we have carried out with Microsoft have allowed us to lay the foundations of a robust technical base and to gain a good understanding of the Azure ecosystem. We naturally wanted to capitalize on these elements."

Human intelligence, pragmatism, and Azure OpenAI Service: The foundations of a successful project

Three months of reflection are necessary to outline the project: "We started cautiously, ensuring not to degrade the customer experience while maintaining human control over the technology," says Sophie Brief. Concretely, the company decides to test the capabilities of generative AI first in its French scope. "Although we operate in about ten countries, France alone represents 60% of our turnover and remains the most advanced in terms of customer engagement," details Jean-Cédric Costa.

As for the use case itself, it focuses on the group's most frequently asked questions (FAQ), already in place on La Redoute's website and mobile application. "The information in our FAQ is generic and simple to process, which is a significant advantage for a first level of experimentation," says Sophie Brief.

Finally, on the generative AI side, the conversational agent is based on Azure OpenAI Service: "The availability of Azure OpenAI Service's Large Language Models and their ease of access bring real added value," emphasizes Jean-Cédric Costa. "A good AI works thanks to qualified, secure, accessible, and efficient data. This is what we benefit from with Microsoft."

Engaging and empowering teams, a lever for acculturation and learning

In parallel, technical and business employees are supported at each key stage of the deployment, both in terms of mastering the technological components and understanding the inherent challenges of customer relations. Workshops led by Microsoft are organized for this purpose, as well as the establishment of a team responsible for internal acculturation.

"It was essential for us to question generative AI and the technological trajectory we want to take in the coming years," explains Jean-Cédric Costa. "It also allowed us to challenge the relevance of the information we provide to customers." From its development to its implementation, the conversational agent has thus been "a powerful vector of learning," adds Sophie Brief.

Concrete gains, surpassing expectations

In just a few months, the solution has delivered greater results than the company expected, providing benefits beyond initial projections. Starting with the level of responses provided by the tool. Indeed, out of the 150,000 customer-assistant AI conversations conducted since the deployment, 60% of the responses provided from La Redoute's mobile app messaging have been effectively handled without human intervention. A strong figure when you know that historically, 50% of the brand's consumers prefer phone contact.

The group's advisors notice a significant improvement in customer support. While about 46% of contact reasons concern order tracking or delivery and 13% request information on the return procedure, "our advisors now intervene directly after a first conversation led by the AI agent on the contact channel of their choice—phone, messaging, or email," observes Sophie Brief. As a result, the topics they handle are more qualified and facilitate the relationship with customers. "The flows directed towards employees decrease, and their quality increases." Not to mention the financial benefits inherent in AI infrastructures and models: "For every euro spent, we save an average of 60 to 70 euros," rejoices Jean-Cédric Costa.

Tomorrow: Continuing to combine progress and quality

Encouraged by these significant gains, the teams are already working on the next steps, namely integrating the agent into the WhatsApp messaging app, developing it over the coming months so that it can provide answers to specific customer questions, well beyond generic FAQ responses, and deploying the solution to other countries where the company operates.

La Redoute also plans to move towards greater personalization of the customer journey. "Thanks to the interactions processed by the AI agent, we have valuable information on our customers' journeys," says Jean-Cédric Costa, "which will give us the means to adapt more quickly to their needs in the future."

In the longer term, the group envisions an era of more complex and autonomous AI agents. A promising innovation but one that cannot succeed without a certain framework. "Whatever our approach, our priority remains to improve our proactivity and responsiveness towards our customers, without compromising on trust or quality."