Emil Frey France, the largest automotive retailer in France, partnered with Kleio to deploy an agentic AI platform on autosphere.fr. The solution understands vehicle projects expressed in natural language, recommends the most relevant vehicles in real time from an inventory of nearly 20,000 used cars, and streamlines the handover to dealerships. The goal is twofold: generate more, and better qualified, online leads for the network, and prepare the native connection of the catalog to AI assistants.

A market that has never been harder to navigate alone

The used-vehicle market has become structurally complex. Electric and hybrid models are arriving in volume, trims and equipment lists have multiplied, financing spans credit, lease-to-own and long-term rental, and trade-in and vehicle taxation shift continually. Each of these has to be explained to every customer. And each used vehicle is unique: one mileage, one model year, one trim, one history, one single unit that leaves the inventory the moment it sells. Meanwhile buyers describe what they want in their own words — a family car, a big boot, nothing that triggers an eco-penalty. Conventional filters and forms can translate neither those vague intentions into concrete criteria, nor match them against stock spread across hundreds of points of sale.

Unifying the data is the hard part

The data required to advise a car buyer properly sits in an unusual number of places: an inventory of nearly 20,000 vehicles refreshed in real time, where every vehicle exists in a single unit; the specifications and equipment of each version; financing grids; tax rules; trade-in data; and the stock held by each individual dealership. Ingesting these heterogeneous sources and making them coherent is where Kleio concentrates its engineering effort. It is that unification work — not a chatbot placed on top of a catalog — that allows AI agents to reason genuinely about the offering.

An AI agent serving buyers and advisors

On autosphere.fr, an AI agent now qualifies each project in natural language: intended use, budget, financing, trade-in and purchase horizon. It translates vague intent into concrete criteria — a family car with a big boot becomes a target load volume, a car with no eco-penalty becomes an emissions threshold — recommends the right vehicles, answers questions on pricing, availability and financing, then passes fully qualified projects to the dealerships. The result: a 20% increase in online leads, of better quality, with advisors engaged only when it makes sense.

Preparing the catalog for AI assistants

The deployment also opens a strategic frontier: connecting the catalog directly to AI assistants such as ChatGPT, Claude and Google AI, where new automotive search intentions increasingly begin. Because Kleio unifies the underlying data, the Autosphere offering is designed to be natively interpretable by those platforms, securing Emil Frey France a leading presence on what is becoming a significant lead generation channel.

A partnership set to expand

This launch is the first step in a broader partnership. Further use cases are already under discussion, including natural-language vehicle comparison and the augmentation of in-dealership advisors by Kleio's AI agents.

“The quality of an AI is determined first and foremost by the data it draws on, far more than by its ability to converse. Used vehicles are the most demanding example of this: thousands of unique cars entering and leaving the inventory every day, with financing and taxation to explain to every customer. Our work consists of ingesting and unifying those sources so that our agents understand that a family car with no eco-penalty corresponds to a very specific set of vehicles in stock. It is that precision which generates better leads for the dealerships and prepares the connection to AI assistants,” explains Philippe Wellens, Co-founder and CEO, Kleio.