LVMH
internal enterprise genAI assistant (marketing content, personalization, research)
MaIA, LVMH's internal genAI assistant built on Google Gemini/Imagen and OpenAI, is used by 40,000 employees for about 1.5 to 2 million requests per month, including generating marketing content and personalized customer messages.
Key points
- Internal enterprise genAI assistant for marketing, creative and retail across the LVMH Maisons.
- MaIA platform built on Google (Gemini, Imagen) and OpenAI (GPT) over a group data foundation.
- 40,000 employees, about 1.5 to 2 million requests per month, business agents at Celine and Tiffany.
- Evidence C, confirmed status: established press citing LVMH and its technology director.
Objective
Equip the marketing, creative and retail teams of the Maisons with a group-wide genAI assistant to speed up content production, customer message personalization and internal research, without exposing the brand universe to consumer chatbots.
The deployment
MaIA is LVMH's internal AI assistant, launched in 2024 and built on a combination of Google models (Gemini, Imagen) and OpenAI (GPT), on top of a group data foundation built with Google. It is used by 40,000 employees and handles on the order of 1.5 to 2 million requests per month. On the marketing and creative side, MaIA is used to generate e-commerce copy, personalized customer messages and mood boards for design teams, in addition to document research and translation. Some Maisons turn the platform into dedicated agents: a retail agent at Celine to answer complex questions from advisors, a customer outreach agent at Tiffany and Co. to write personalized messages. LVMH deliberately keeps these tools internal and does not put a chatbot on its sales sites.
Results Proof C
Two established press outlets (PYMNTS, WWD via Yahoo Finance) naming LVMH and its technology director, with concordant usage figures (40,000 employees, 1.5 to 2M requests/month). No direct financial result attached to MaIA.
How it works
Inferred typical approachThe internal detail is not public. Here is a proven approach that leads to the same result, to adapt to your stack.
The stack in detail
- llm Google Gemini language model for text generation, translation and internal research
- llm Google Imagen image generation, notably for the design teams' mood boards
- llm OpenAI GPT second text generation foundation; MaIA routes the request to the model suited to the task
- infra Socle data groupe sur Google Cloud grounds the generation on the group's data and tone, with isolation from third-party models and access logging
- plateforme MaIA (plateforme interne LVMH) group-wide genAI assistant used by 40,000 employees (1.5 to 2M requests per month), turned into business agents by the Maisons (Celine, Tiffany)
How it runs, concretely
For ops teams-
1Business request marketing / creative
A marketing or creative employee asks for copy, an image, a translation or research.
-
2Multi-model processing AI / platform
MaIA routes to Gemini, Imagen or GPT depending on the task, anchored on the group's data and tone.
-
3Business agent Maison team
A Maison turns MaIA into a dedicated agent (Celine retail, Tiffany outreach) for a targeted use.
-
4Human control marketing / advisor
The employee reviews and validates the content before distribution; the sales sites remain without a chatbot.
Governed access to the group data foundation and brand data. Without a connection to internal data and brand tone, generation stays generic and unusable in luxury.
How your customers perceive this type of use
Sourced studiesUn ecart net separe les annonceurs des consommateurs : 77% des annonceurs voient l'IA positivement contre 38% des consommateurs (Yahoo/Publicis, 2024). Les mesures implicites confirment le rejet declare : en EEG, les pubs generees par IA produisent une activation memorielle plus faible que les pubs traditionnelles et sont decrites comme agacantes, ennuyeuses et confuses (NIQ, 2024). La disclosure a un effet ambivalent : elle augmente fortement la confiance quand elle est remarquee (Yahoo/Publicis), mais 27% des jeunes consommateurs disent faire moins confiance a une entreprise dont la pub est creee par IA (IAB, 2024).
Acceptance conditions
- Une disclosure visible : quand la mention IA est remarquee, la confiance globale envers l'entreprise augmente de 96% (Yahoo/Publicis 2024)
- Une qualite visuelle suffisante : les visuels IA de basse qualite augmentent l'effort cognitif et distraient du message (NIQ 2024)
Red lines
- Le contenu IA non declare puis identifie : 72% des consommateurs disent que l'IA rend l'authenticite difficile a etablir (Yahoo/Publicis 2024) et les marques utilisant des pubs IA sont plus souvent jugees inauthentiques ou non ethiques par les consommateurs que par les dirigeants (IAB 2024)
- Les mannequins et personnes generes par IA : 46% des consommateurs n'en veulent pas dans la publicite, l'inquietude premiere etant les standards de beaute irrealistes (Attest 2025)
Sources: Yahoo / Publicis Media (terrain Ebco) 2024 · IAB (avec Attest) 2024 · NIQ (NielsenIQ) 2024 · Attest 2025
How to replicate
Inference, not sourcedData prerequisites
- governed enterprise data foundation
- reference set for brand tone and universe
- access governance for sensitive data (HR, customer)
Org prerequisites
- IT department to run the central platform
- enterprise genAI licenses with data isolation
- a liaison in each brand to build the business agents
Possible stack
- Google Gemini / Vertex AI
- OpenAI ChatGPT Enterprise / Azure OpenAI
- image generation (Imagen, DALL-E)
- agent orchestration
The plan, step by step
- Step 1Frame the data governance and two priority marketing use cases (e-commerce copy, mood boards)Deliverable: v1 scope + access rules for sensitive data (HR, customer)
- Step 2Contract the models in enterprise versions with data isolation and loggingDeliverable: Isolated multi-model API access, validated by security
- Step 3Build the cross-functional assistant anchored on the data foundation and the brand tone referenceDeliverable: v1 assistant opened to a first group of marketing and creative users
- Step 4Deploy more broadly, train the teams and track adoptionDeliverable: Requests and adoption dashboard by function
- Step 5Turn it into business agents per brand (retail support, customer outreach), with human review before distributionDeliverable: 1 to 2 dedicated agents in production in the pilot brands
First step: Open an enterprise genAI assistant anchored on internal data and brand tone, with two priority marketing use cases (copy, mood boards).
The same pattern, proven in other industries
Adobe
in-house genAI production of marketing content at scale
Atlassian
internal genAI agents for account targeting and marketing content production
BMW
advertising and experiential creation with generative AI
Sources
- S1 LVMH Deploys AI Tools Across Operation, Seeking Efficiency and Customer Retention Established press archive pending
- S2 LVMH and Google Executives Talk Agentic AI, Cybersecurity and Navigating Volatility (WWD) Established press archive pending
An error, newer info, a source?
This page lives on its accuracy. If a figure has moved, if the deployment has changed, or if you have a higher-quality source, tell us. Every sourced correction is verified before publication.