Verified
L'Oréal
Piloting- Beauty
- 95,000+ staff
- €1.5B tech investment (2025)
- L'OréalGPT, built in-house
- 56,000+ with access
- AI content cost down 40%
L'Oréal deploys AI through tools it built in-house, chiefly the L'OréalGPT employee assistant, the CreAItech content studio and the BETiq media planner. The tools run models from Google, OpenAI, Anthropic, NVIDIA and IBM. More than 56,000 L'Oréal employees had L'OréalGPT access and 73,000 had taken generative AI training by Apr 2026. CEO Nicolas Hieronimus said in Feb 2026 that CreAItech had cut the cost of each piece of content by 40%, in a year when L'Oréal invested a record €1.5 billion in tech.
loreal.comTheir approach to AI
- The platform
- L'Oréal builds its own AI tools on bought models: L'OréalGPT for employees, CreAItech for marketing content, BETiq for media spend and a Beauty Tech Agentic Platform on Google Cloud. The tools all draw on L'Oréal's Beauty Tech Data Platform (Apr 2026).
- The models
- L'Oréal picks models per tool: Google Gemini, Imagen and Veo in CreAItech, and OpenAI's GPT-4o in Beauty Genius and GPT-Rosalind in research. L'Oréal also uses Anthropic Claude for analytics agents, NVIDIA software and a formulation model built with IBM (2025 to 2026).
- The reach
- More than 56,000 of L'Oréal's more than 95,000 employees had L'OréalGPT access in Apr 2026, and about 21,000 used it daily. By then, 73,000 L'Oréal employees had taken the Gen AI for All course.
- The catch
- L'Oréal's 2 published AI returns, a 40% lower cost per piece of content and 10% to 15% more productive media spend, both sit in marketing. L'Oréal gives no firm-level AI figure and has said more will come at an AI-focused capital markets day.
Snapshot
- Employees
- 95,000+
- Company description, 2026
- Technology investment
- €1.5B (2025)
- A record year; R&I spending was €1.4B, 3.1% of sales
- AI/ML headcount
- Not disclosed
- 8,000+ digital, tech and data talents; 2,000 tech jobs planned in Hyderabad by 2030
- Who owns AI
- Samuel du Retail
- General Manager AI, Data and Shared Services; leads the AI transformation team created by 2026
- Platform
- L'OréalGPT, CreAItech, BETiq
- Built in-house; Beauty Tech Agentic Platform on Google Cloud
- Models
- Google, OpenAI, Anthropic, NVIDIA, IBM
- Chosen per tool; L'Oréal has not said which models run L'OréalGPT
- Reach
- 56,000+ with L'OréalGPT access
- Apr 2026; about 21,000 daily users and 22,000+ personal AI companions
- Use cases
- Marketing, research, consumer care
- Content, media spend, formulation, beauty advice, service agents
Key findings
Nicolas Hieronimus said on Feb 13, 2026 that L'Oréal's CreAItech studio makes about 500,000 pieces of content a month. Hieronimus said the cost of one piece of content is down 40%, the clearest return L'Oréal credits to AI.
L'Oréal's BETiq tool, which uses AI to allocate advertising and promotion spend, generates 10% to 15% productivity in each country's first year, Hieronimus said on Feb 20, 2026. BETiq was set to cover 10 countries and more than half of L'Oréal's optimizable advertising and promotion spend in 2026.
L'Oréal invested a record €1.5 billion in tech in 2025. Hieronimus told investors in Feb 2026 that a newly created AI transformation team is running AI use cases at scale on top of a single IT backbone.
Samuel du Retail told shareholders on Apr 24, 2026 that 73,000 L'Oréal employees had been trained in generative AI, more than 56,000 had access to L'OréalGPT and more than 22,000 personal AI companions had been built. L'Oréal's L'OréalGPT counts each use a different measure: 32,000 monthly users in 2024, 71,000 users in 2025 in the registration document, and about 21,000 daily users in Apr 2026; an undated Anthropic case study separately counts 44,000 monthly unique users of L'Oréal's internal AI platform.
L'Oréal named OpenAI a foundational partner on Jun 17, 2026, adding ChatGPT product discovery, a Maybelline virtual try-on inside ChatGPT and the GPT-Rosalind life sciences model to work already running with Google Cloud, NVIDIA, IBM, Microsoft and Anthropic.
Where L'Oréal sits
- Stage
- Piloting
L'Oréal has reported returns it credits to AI in one function, marketing, which places it at Piloting. CEO Nicolas Hieronimus said on Feb 13, 2026 that the cost of one piece of content made in the CreAItech studio is down 40%. On Feb 20, 2026, Hieronimus said the BETiq media tool generates 10% to 15% productivity on advertising and promotion spend in each country's first year. On rollout, Samuel du Retail, General Manager AI, Data and Shared Services, told the Apr 24, 2026 annual meeting that more than 56,000 L'Oréal employees have access to L'OréalGPT and 73,000 have been trained in generative AI. Du Retail was also reported saying 21,000 employees use L'OréalGPT daily. L'Oréal falls short of Scaling because both returns come from marketing and because it has published no firm-level AI figure. Hieronimus says L'Oréal's AI use cases represent significant productivity on several P&L lines but gives no number. CFO Christophe Babule's expected 20 to 30 basis points of yearly improvement in selling, general and administrative costs is credited to IT transformation and AI together.
How the stages are defined - Ownership
- Samuel du Retail, General Manager AI, Data and Shared Services; a new AI transformation team
- Platform
- Hybrid: L'OréalGPT, CreAItech and BETiq built in-house on bought models
- Models
- Google, OpenAI, Anthropic, NVIDIA and IBM, chosen per tool
- Hosting
- Google Cloud for the agent platform and CreAItech models, Microsoft Azure for Beauty Genius and Noli, NVIDIA AI Enterprise software; own data centers not disclosed
- Reach
- 56,000+ with L'OréalGPT access, about 21,000 daily (Apr 2026)
How they deploy
Consumers, business functions and employees
L'Oréal organizes its AI program around serving consumers, serving business functions and serving employees, Samuel du Retail told the Apr 2026 annual meeting. The aim is to become One L'Oréal powered by AI. Nicolas Hieronimus told investors in Feb 2026 that AI is at the core of L'Oréal's new playbook and that a newly created AI transformation team is putting use cases into production at scale.
Build the tools, mix the models
L'Oréal designs its own tools and plugs outside models into them. CreAItech, built by L'Oréal, combines models from Google, Adobe, NVIDIA, Seedance and OMI. Chief Digital and Marketing Officer Asmita Dubey said in Jun 2026 that L'Oréal added OpenAI's image model because models come with different strengths. Group CIO Etienne Bertin said in Apr 2026 that Google Cloud gives L'Oréal's agent platform the flexibility to use several large language models. Thomas Menard, head of the agentic platform, said L'Oréal picks models through automated evaluations, such as having one model grade another's answers.
IT backbone before AI
L'Oréal treats its IT overhaul as the base for AI. Hieronimus called the deployment of the One L'Oréal IT backbone the first and necessary step for AI in Feb 2026, and said in Apr 2026 that L'Oréal had harmonized 20 IT systems. Consumer Goods Technology reported that L'Oréal expected about 60% of its IT transformation to be complete by the end of 2026.
Research first
CFO Christophe Babule said in Sep 2026 that research is by far the first area where L'Oréal focuses AI. Babule said that is why L'Oréal partners with IBM, NVIDIA and OpenAI, with the goal of bringing new formulas 2, 3 or 5 times faster. Hieronimus said in Jul 2026 that Research & Innovation is the function AI has augmented most.
Visible inside the chatbots
L'Oréal wants its brands to be the source large language models draw on when consumers ask about beauty. Hieronimus said in Feb 2026 that L'Oréal aims to be the reference source of beauty knowledge for these models through partnerships with the major players. A generative engine optimization task force ran 30 pilot tests in 2025, and L'Oréal has run ads on ChatGPT with CeraVe, SkinCeuticals and Garnier since Apr 2026.
- Deployment model
- L'Oréal builds AI tools centrally and puts them to work in the business functions. Samuel du Retail's AI, Data and Shared Services organization and the AI transformation team set the program. Group IT under CIO Etienne Bertin runs the Beauty Tech Agentic Platform on Google Cloud, and Asmita Dubey's digital and marketing organization owns CreAItech and BETiq. Research & Innovation, under Deputy CEO Barbara Lavernos, runs the formulation models with IBM and NVIDIA. Countries too small for a full BETiq deployment get BETiq Scaler, an AI tool that applies lessons from the larger markets (Feb 2026).
- Model providers
- L'Oréal uses several model providers at once. CreAItech runs Google's Gemini, Imagen and Veo models alongside Adobe, NVIDIA, Seedance and OMI tools and, from Jun 2026, OpenAI's latest image model. Beauty Genius runs on GPT-4o through Microsoft's Azure OpenAI Service. An undated Anthropic case study says Claude coordinates more than 15 specialized agents for conversational analytics on L'Oréal's internal AI platform, and that engineers use GitHub Copilot and Claude Code. Research uses a Formulation Foundation Model built with IBM, NVIDIA's ALCHEMI framework and OpenAI's GPT-Rosalind. L'Oréal has not said which models run L'OréalGPT.
- Data and infrastructure
- L'Oréal says it holds the richest beauty database in the world, with more than 1.8 million formulas used to train its IBM model. Hieronimus cited 16,000 terabytes of data in Jun 2025. L'Oréal's agents, built on Google Cloud, reach the Beauty Tech Data Platform and core operational applications through the Model Context Protocol, according to Etienne Bertin (Apr 2026). Beauty Genius runs on Microsoft Azure with interactions kept in a dedicated L'Oréal environment. Noli's AI Refinery, built with NVIDIA and Accenture, is available on Microsoft Azure (Jun 2025). Consumer care agents run in Salesforce Service Cloud and draw on Salesforce Data 360 and order data from SAP S/4HANA (Oct 2026). Hieronimus said in Apr 2026 that L'Oréal works with data center partners that prioritize low-carbon energy. L'Oréal has not said whether any models run in its own data centers.
Platforms and internal tools
L'OréalGPT and AI companionsAbout 2024
L'OréalGPT is L'Oréal's secure in-house conversational assistant for employees, which Samuel du Retail said launched about 2 years before Apr 2026. Employees build personal AI companions on L'OréalGPT for writing, brainstorming, project analysis, translation and planning. More than 22,000 individual companions and 254 broader companions under business governance existed in Apr 2026.
- Users
- 56,000+ with access and about 21,000 daily (Apr 2026); 32,000 monthly in 2024
CreAItech2024
CreAItech is L'Oréal's in-house generative AI content platform, which creates images, video, sound and 3D digital twins of products for brands such as L'Oréal Paris, Yves Saint Laurent Beauty and Maybelline. L'Oréal began building CreAItech in 2024 and rolled it out in 2025. Nicolas Hieronimus said in Jul 2026 that CreAItech makes more than 500,000 pieces of content a month. CreAItech has a built-in CO2 estimator for each creation.
- Users
- 4,500 (2025); thousands of users across the group
- Models
- Google Gemini, Imagen and Veo; Adobe; NVIDIA; Seedance; OMI; OpenAI from Jun 2026
BETiq
BETiq is L'Oréal's proprietary AI and predictive modeling tool for allocating advertising and promotion spend across more than 50 touchpoints. BETiq ran in 9 countries in 2025 and was planned for 10 in 2026, covering more than half of L'Oréal's optimizable advertising and promotion spend. BETiq Scaler extends BETiq's lessons to smaller countries.
- Users
- 1,000+ people trained (2025)
Beauty Tech Agentic PlatformBy Apr 2026
The Beauty Tech Agentic Platform is L'Oréal's proprietary platform for AI agents, built on Google Cloud with the Agent Development Kit, according to Group CIO Etienne Bertin. Bertin said the platform connects through the Model Context Protocol to the Beauty Tech Data Platform and core operational applications.
Beauty GeniusUnveiled Jan 2024
Beauty Genius is L'Oréal Paris's AI beauty assistant, unveiled at CES 2024, which gives personalized diagnosis, product recommendations and where-to-buy advice in the United States. Beauty Genius runs on GPT-4o through Azure OpenAI Service and had generated more than 1.1 million conversations by the 2025 annual report.
- Models
- OpenAI GPT-4o via Microsoft Azure
Formulation Foundation ModelAnnounced Jan 2025
The Formulation Foundation Model is a custom AI model L'Oréal Research & Innovation is building with IBM to help researchers design formulas and switch to sustainable raw materials. The model is trained on more than 1.8 million formulas, and L'Oréal called the early results promising in its 2025 annual report and said a next iteration is under way.
- Models
- IBM
NoliBy Jun 2025
Noli is an AI-powered multi-brand beauty marketplace founded and backed by L'Oréal that matches each shopper's beauty profile to products, using more than 1 million skin data points. Noli's AI Refinery was built with NVIDIA and Accenture on Microsoft Azure (Jun 2025).
Vendors and partners
Google Cloud
Cloud for the Beauty Tech Agentic Platform; Gemini, Imagen and Veo models in CreAItech · Group CIO Etienne Bertin said in Apr 2026 that L'Oréal built its agentic platform on Google Cloud for resilience and the flexibility to use several large language models.
OpenAIJun 2026 partnership
Foundational partner for consumer journeys and research; GPT-4o behind Beauty Genius · The Jun 17, 2026 agreement covers ChatGPT product discovery, a Maybelline virtual try-on in ChatGPT, a ChatGPT ad pilot, GPT-Rosalind for skin microbiome research and OpenAI's model in CreAItech.
Microsoft
Azure OpenAI Service hosts Beauty Genius; Azure hosts Noli's AI Refinery; GitHub Copilot for engineers · A Microsoft case study says L'Oréal moved Beauty Genius to a mix of provisioned throughput and pay-as-you-go capacity to keep latency under 5 seconds.
NVIDIAJun 2025
NVIDIA AI Enterprise for CreAItech's 3D rendering; ALCHEMI for computational chemistry in research · L'Oréal expanded the partnership on Mar 17, 2026 to predict how molecules perform at atomic scale, starting with photoprotection and skin tone.
IBMJan 2025
Co-developer of the Formulation Foundation Model
Anthropic
Claude models for analytics agents on L'Oréal's internal AI platform; Claude Code in testing · Anthropic's undated case study says 30 to 50 L'Oréal developers were testing Claude Code.
Salesforce
Agentforce in Service Cloud for consumer care advisers · Deployed across 25 brands in North America by Oct 2026, with Europe next.
Adobe
Models and tools in CreAItech
AccentureJun 2025
Built Noli's AI Refinery with NVIDIA
Leadership
AI, data and platform leaders
Executives
AI org map
Central AI, data and IT functions build shared tools; marketing and research run their own AI programs
L'Oréal's AI program sits with Samuel du Retail, General Manager AI, Data and Shared Services, and an AI transformation team that Nicolas Hieronimus described as newly created in Feb 2026. Deputy CEO Barbara Lavernos oversees Research, Innovation and Technology, Chief Digital and Marketing Officer Asmita Dubey owns the marketing AI tools, and Group CIO Etienne Bertin runs the agentic platform. Lavernos, Dubey and Chief Innovation & Prospective Officer Delphine Viguier-Hovasse sit on the executive committee. Samuel du Retail and Etienne Bertin do not appear on the committee's Mar 1, 2026 list.
L'Oréal's AI work draws on more than 8,000 digital, tech and data talents and more than 4,000 Research & Innovation scientists (2026). Samuel du Retail's AI, Data and Shared Services organization and a newly created AI transformation team set the program (Feb 2026). L'Oréal has not published a count of AI and machine learning specialists. L'Oréal is adding a Beauty Tech hub in Hyderabad, India, with about 2,000 tech jobs planned by 2030 to develop AI beauty services for the whole group (Jan 2026).
Reporting lines
Chief Executive Officer
Nicolas Hieronimus
AI, Data and Shared Services
Leads the AI program and presented it to shareholders in Apr 2026; reporting line not published
Research, Innovation and Technology
Barbara Lavernos, Deputy CEO
4,000+ scientists; formulation models with IBM and NVIDIA
Group IT
Etienne Bertin, Group CIO
Runs the Beauty Tech Agentic Platform; line inferred from Lavernos's technology remit, not published
Digital and Marketing
Asmita Dubey, Chief Digital and Marketing Officer
Owns CreAItech and BETiq
Innovation and Prospective
Finance
Christophe Babule, CFO
Uses AI in accounting, treasury and financial control (Feb 2026)
L'Oréal has not published to whom Samuel du Retail or Etienne Bertin report. L'Oréal's own materials describe an AI transformation team without saying where it sits.
Where the AI workforce sits
Digital, tech and data teams
Estimate: 8,000+ digital, tech and data talents out of about 12,000 technical staff, counting the 4,000+ scientists; L'Oréal does not say how many of either work on AI.
~65%Research & Innovation scientists
Estimate: 4,000+ scientists who use AI formulation and screening tools; how many build AI models is not disclosed.
~35%
Named people
Use cases
L'Oréal runs AI across marketing content, media spend, research and formulation, consumer advice and consumer care, with employee tools built on L'OréalGPT. L'Oréal has published returns it credits to AI for 2 marketing use cases. In the CreAItech studio, the cost of one piece of content is down 40% (Feb 2026). The BETiq media tool generates 10% to 15% productivity on advertising and promotion spend in each country's first year (Feb 2026). L'Oréal reports its research gains as speed and volume, not cost, and gives no firm-level AI figure. Nicolas Hieronimus said in Jul 2026 that more would come at an AI-focused investor event in Paris later in 2026.
| Use case | Area | Status | ROI |
|---|---|---|---|
| CreAItech content production CreAItech generates on-brand images, video and 3D product renderings for social media, e-commerce and campaigns, at more than 500,000 pieces of content a month across L'Oréal's brands (Jul 2026). | Marketing | In production | Cost of one piece of content down 40% (Feb 2026)Source |
| BETiq media and promotion allocation BETiq uses AI and predictive modeling to allocate advertising and promotion spend across more than 50 touchpoints. BETiq ran in 9 countries in 2025 and was planned to cover more than half of L'Oréal's optimizable advertising and promotion spend across 10 countries in 2026. | Marketing | In production | 10% to 15% productivity on advertising and promotion spend in each country's first year (Feb 2026)Source |
| Revenue growth management L'Oréal uses its data and AI to set prices in its largest markets, a program CFO Christophe Babule named in Feb 2026 among the group's main AI initiatives. | Marketing | In production | |
| Marketing studies agent An AI document search agent lets more than 10,000 research and marketing staff search more than 100,000 proprietary marketing studies from 70 countries and summarizes consumer trends for international teams (Apr 2026). | Marketing | In production | |
| Brand visibility in AI assistants L'Oréal supplies product information directly to OpenAI for ChatGPT answers in the United States and ran 30 generative engine optimization pilot tests in 2025. L'Oréal has advertised CeraVe, SkinCeuticals and Garnier on ChatGPT since Apr 2026 as part of OpenAI's ad pilot. | Marketing | Pilot | |
| Beauty Genius L'Oréal Paris's Beauty Genius assistant gives US consumers personalized diagnosis, recommendations and where-to-buy advice, and had generated more than 1.1 million conversations by the 2025 annual report. | Consumer | In production | |
| Maybelline virtual try-on in ChatGPT Maybelline New York is bringing its ModiFace-powered makeup virtual try-on into ChatGPT, under the Jun 2026 OpenAI agreement. | Consumer | Announced | |
| Noli marketplace Noli, a multi-brand marketplace backed by L'Oréal, decodes each shopper's beauty profile with AI diagnostics and recommends products from many brands. | Consumer | In production | |
| AI agents for consumer care advisers L'Oréal's consumer care agents detect what a consumer is asking, gather the answer in the right language and draft the reply. Samuel du Retail said the agents make responses 3 times faster and cut advisers' administrative time by up to 80% (Apr 2026). In North America, Salesforce Agentforce runs across 25 brands for more than 300 advisers a day and cuts case handling time by up to 64%, according to Salesforce (Oct 2026). | Consumer care | In production | |
| Shelf checks by visual recognition Visual recognition agents help L'Oréal sales representatives check in-store product availability and spot likely stock shortages, cutting verification time by more than 50% with 99% reliability, du Retail said in Apr 2026. | Sales | In production | |
| K-SCAN salon diagnostics K-SCAN is an AI-powered camera that gives salon clients personalized hair recommendations, deployed in 10,000 salons worldwide (Feb 2026). L'Oréal says K-SCAN raises salon sales by an average of 23%, a gain for its salon customers. | Professional Products | In production | |
| Hair digital twins and molecule screening L'Oréal's digital hair reacts to active ingredients like its physical twin. Samuel du Retail said in Apr 2026 that over 4 years L'Oréal tripled the number of molecules tested and cut their development time by a factor of 4. Nicolas Hieronimus said in Feb 2026 that AI had saved up to 40% of time to market. | Research & Innovation | In production | |
| Formulation Foundation Model with IBM The Formulation Foundation Model, trained on more than 1.8 million formulas, helps L'Oréal researchers design new textures, tune formula performance and use sustainable raw materials. L'Oréal described early results as promising in its 2025 annual report. | Research & Innovation | Pilot | |
| Molecular simulation with NVIDIA ALCHEMI L'Oréal is building a skincare AI engine on NVIDIA's ALCHEMI framework to predict how molecules perform at atomic scale, starting with photoprotection and skin tone management. L'Oréal says the engine makes discovery 100 times faster than traditional methods (Mar 2026). | Research & Innovation | Pilot | |
| Skin microbiome mapping with GPT-Rosalind L'Oréal is using OpenAI's GPT-Rosalind life sciences model to map the skin microbiome and identify beneficial bacteria for new skincare, starting with La Roche-Posay (Jun 2026). | Research & Innovation | Pilot | |
| L'OréalGPT and personal AI companions L'OréalGPT gives employees a secure assistant for daily work, with more than 1 million conversations a week and more than 22,000 personal companions built for writing, translation and planning (Apr 2026). | Employees | In production | |
| Conversational analytics agents L'Oréal employees ask questions of company data in natural language, and Claude coordinates more than 15 specialized agents to return answers and charts from the Beauty Tech Data Platform, according to Anthropic. The agents replace a custom dashboard for each question. | Employees | In production | |
| Finance productivity Christophe Babule said in Feb 2026 that AI tools already bring productivity to L'Oréal's accounting, treasury and financial control, with gains to come year after year rather than in one step. | Finance | In production | |
| AI coding tools L'Oréal engineers use GitHub Copilot and Claude Code across the software development lifecycle, with 30 to 50 developers testing Claude Code, according to Anthropic. | IT | Pilot |
Adoption
L'Oréal spread generative AI to employees through training first. The Gen AI for All course reached 42,000 staff in 2024, more than 65,000 by Dec 2025 and 73,000 by Apr 2026, and was extended to plants, distribution centers and points of sale. AmplifAI self-assessments began in May 2026, with a goal of 40,000 completions by the end of 2026. L'Oréal reports L'OréalGPT use under several measures that should not be merged. The 2024 figure was 32,000 monthly users; the 2025 annual report gives more than 60,000 users, and the registration document 71,000 users in 2025. In Apr 2026, more than 56,000 employees had access and about 21,000 used L'OréalGPT daily. An undated Anthropic case study separately counts 44,000 monthly unique users and 15,000 daily unique users of L'Oréal's internal AI platform. L'Oréal had more than 95,000 employees in 2026, so daily L'OréalGPT users were under a quarter of staff in Apr 2026.
By the numbers
Rollout, first pilot to today
2023
L'Oréal added a Responsible Framework for Trustworthy AI to its Code of Ethics.
Jan 2024
Nicolas Hieronimus unveiled the Beauty Genius AI assistant at CES 2024 and said L'Oréal held 10 petabytes of data on its data platform.
2024
42,000 L'Oréal employees took the Gen AI for All course, and L'OréalGPT reached 32,000 monthly users.
Jan 2025
L'Oréal and IBM announced a custom AI foundation model for cosmetic formulation.
Jun 2025
L'Oréal began working with NVIDIA AI Enterprise to scale CreAItech's 3D product rendering, and Noli unveiled an AI Refinery built with NVIDIA and Accenture.
Dec 2025
More than 65,000 L'Oréal employees had been trained on generative AI and more than 60,000 used L'OréalGPT, up 25% on the year, according to the 2025 annual report.
Feb 2026
Hieronimus said L'Oréal had created an AI transformation team after a record €1.5 billion tech investment in 2025, and that BETiq would cover 10 countries in 2026.
Apr 2026
Samuel du Retail told shareholders that 73,000 employees had been trained, more than 56,000 had L'OréalGPT access and more than 22,000 AI companions had been built.
Jun 2026
L'Oréal named OpenAI a foundational partner for consumer journeys, research and marketing content.
Oct 2026
Salesforce reported Agentforce running for L'Oréal consumer care advisers across 25 brands in North America, with all eligible advisers using it.
Economics & controls
Spend and ROI
Technology budget
L'Oréal invested a record €1.5 billion in tech in 2025, covering the One L'Oréal IT backbone that Hieronimus called the base for its AI work (Feb 2026). L'Oréal spent €1.4 billion on Research & Innovation in 2025, 3.1% of sales. L'Oréal also spent €14.2 billion on advertising and promotion, 32.2% of sales, the spend that BETiq and CreAItech serve. L'Oréal committed an initial investment of more than 35 billion rupees, about $383 million, to a Beauty Tech hub in Hyderabad for AI beauty work (Jan 2026).
What AI itself costs
L'Oréal does not break out what AI costs. A Microsoft case study says L'Oréal moved Beauty Genius to a hybrid of provisioned and pay-as-you-go Azure OpenAI capacity to control cost and latency. Hieronimus said in Apr 2026 that L'Oréal measures the energy use of its AI and prefers the most energy-efficient conversational models.
Reported ROI
L'Oréal credits AI with 2 returns, both in marketing. Hieronimus said on Feb 13, 2026 that the cost of one piece of CreAItech content is down 40%. On Feb 20, 2026, Hieronimus said BETiq generates 10% to 15% productivity on advertising and promotion spend in each country's first year. Christophe Babule said BETiq brings a better return at least in year one in every country where it was rolled out. Babule also said in Jul 2025 that advertising and promotion fell 20 basis points to 31.9% of sales in the first half of 2025 partly thanks to BETiq, without separating BETiq from other factors. Hieronimus said in Oct 2026 that AI lets L'Oréal do more for the same marketing money, or spend less if it chose. L'Oréal gives no firm-level AI figure. Hieronimus said in Feb 2026 that L'Oréal's AI use cases represent significant productivity on multiple P&L lines, but gave no number.
Productivity markers
L'Oréal reports most AI gains as speed or volume, not cost. L'Oréal's research teams tested more molecules in the 12 months to Feb 2026 than in the previous 5 years and saved up to 40% of time to market, according to Hieronimus. Consumer care advisers spend up to 80% less time on administrative tasks, and sales representatives verify shelf availability more than 50% faster, according to Samuel du Retail (Apr 2026). Christophe Babule said in Jul 2026 that L'Oréal could expect around 20 to 30 basis points a year of improvement in selling, general and administrative costs from investing in IT transformation and AI. Babule did not say how much comes from each.
Headcount
L'Oréal had more than 95,000 employees in 2026, including more than 8,000 digital, tech and data talents and more than 4,000 scientists. Hieronimus said in Apr 2026 that AI takes on repetitive tasks such as data compilation so employees can do higher-value work, and that AI is not meant to replace them. Alberto Rodrigues of L'Oréal USA said the consumer care agents augment advisers rather than replace them (Oct 2026). L'Oréal has published no AI-related job cuts and plans about 2,000 tech jobs in Hyderabad by 2030.
Governance and risk
Trustworthy AI framework
L'Oréal added a Responsible Framework for Trustworthy AI to its Code of Ethics in 2023 and builds ethics into the Gen AI for All course. By Dec 31, 2025, 80% of L'Oréal's Data & Tech staff had taken a dedicated Trustworthy AI course, according to the 2025 registration document. L'Oréal lists the responsible use of AI among its ethics risks.
No AI-generated people in claims
L'Oréal's CreAItech policy bars AI-generated lifelike faces, bodies, hair or skin from supporting or enhancing product benefits. L'Oréal also gives staff guidelines on creating and distributing AI-generated content to manage intellectual property risk, such as uncertain ownership of AI-made works.
Bias and energy
Hieronimus said in Apr 2026 that L'Oréal works to address algorithmic bias so its systems work for all skin tones and hair types. Hieronimus said L'Oréal also measures the energy its technologies use, prefers energy-efficient models and uses data center partners that prioritize low-carbon energy. CreAItech shows each creator an estimate of the CO2 behind their work.
Data boundary and agent access
L'OréalGPT is a secure internal tool for employees, and Beauty Genius keeps consumer conversations in a dedicated L'Oréal environment on Azure. L'Oréal's conversational analytics agents manage user identity and access controls when they query the Beauty Tech Data Platform. For consumer care agents, L'Oréal maps the exact data each agent may pull so sensitive systems stay closed off, and the same map governs every agent L'Oréal builds, according to Salesforce (Oct 2026).
Board and pay
L'Oréal's 2025 registration document lists L'OréalGPT use, CreAItech, BETiq and the AI and Beauty Tech transformation among the non-financial achievements assessed for Nicolas Hieronimus's 2025 pay. Hieronimus represented L'Oréal in the EU AI Champions Initiative launched on Feb 10, 2025.
Recent signals
Oct 22, 2026
Nicolas Hieronimus said on the third-quarter call that AI helps L'Oréal allocate media through BETiq and generate content variations, letting it do more for the same money. Hieronimus also said L'Oréal will cover AI at its capital markets day.
Oct 20, 2026
Salesforce published L'Oréal's use of Agentforce for consumer care advisers across 25 brands in North America, with more than 300 daily users and Europe next.
Oct 2026
Berenberg raised its L'Oréal price target to 480 euros, saying AI-powered search is starting to favor L'Oréal in beauty discovery, and named a Nov 30, 2026 AI-focused investor day as the next test.
Sep 8, 2026
CFO Christophe Babule told the Barclays consumer conference that research is by far the first area of L'Oréal's AI focus, with IBM, NVIDIA and OpenAI as partners.
Jul 30, 2026
Hieronimus said CreAItech makes more than 500,000 pieces of content a month and that he had taken L'Oréal's executive committee to Silicon Valley on an AI learning trip. Hieronimus also said investors would get an AI session in Paris later in 2026.
Jun 22, 2026
Chief Innovation & Prospective Officer Delphine Viguier-Hovasse told Fortune that AI lets L'Oréal test a molecule in 3 months, where 20 molecules on hair used to take several years.
Jun 17, 2026
L'Oréal and OpenAI announced a collaboration at VivaTech covering ChatGPT product discovery, a Maybelline try-on in ChatGPT, a ChatGPT ad pilot, GPT-Rosalind in skin research and OpenAI's model in CreAItech.
Apr 24, 2026
Samuel du Retail presented L'Oréal's AI strategy at the annual meeting: 73,000 employees trained, more than 56,000 with L'OréalGPT access and more than 22,000 AI companions.
Apr 22, 2026
Google Cloud quoted Group CIO Etienne Bertin on L'Oréal's Beauty Tech Agentic Platform, built on Google Cloud and connected to the Beauty Tech Data Platform through the Model Context Protocol.
Mar 17, 2026
L'Oréal expanded its NVIDIA partnership into research, building a skincare AI engine on NVIDIA's ALCHEMI framework for molecular simulation.
Feb 20, 2026
Hieronimus told the CAGNY conference that BETiq generates 10% to 15% productivity on advertising and promotion spend in each country's first year.
Feb 13, 2026
Hieronimus said L'Oréal invested a record €1.5 billion in tech in 2025, created an AI transformation team, and cut the cost of one piece of CreAItech content by 40%.
Jan 21, 2026
L'Oréal said it will set up a Beauty Tech hub in Hyderabad, India, with an initial investment of more than 35 billion rupees and about 2,000 tech jobs by 2030, to build AI beauty services.
Dec 3, 2025
L'Oréal USA's chief executive told the capital markets day that almost 80% of the US business's advertising and promotion spend runs through BETiq and that AI is cutting the cost per piece of content sharply.
Quick answers
- How does L'Oréal deploy AI?
- L'Oréal deploys AI through tools it built in-house, chiefly the L'OréalGPT employee assistant, the CreAItech content studio and the BETiq media planner. The tools run models from Google, OpenAI, Anthropic, NVIDIA and IBM. More than 56,000 L'Oréal employees had L'OréalGPT access and 73,000 had taken generative AI training by Apr 2026. CEO Nicolas Hieronimus said in Feb 2026 that CreAItech had cut the cost of each piece of content by 40%, in a year when L'Oréal invested a record €1.5 billion in tech.
- What does L'Oréal use AI for?
- L'Oréal runs AI across marketing content, media spend, research and formulation, consumer advice and consumer care, with employee tools built on L'OréalGPT. L'Oréal has published returns it credits to AI for 2 marketing use cases. In the CreAItech studio, the cost of one piece of content is down 40% (Feb 2026). The BETiq media tool generates 10% to 15% productivity on advertising and promotion spend in each country's first year (Feb 2026). L'Oréal reports its research gains as speed and volume, not cost, and gives no firm-level AI figure. Nicolas Hieronimus said in Jul 2026 that more would come at an AI-focused investor event in Paris later in 2026.
- Who leads AI at L'Oréal?
- AI at L'Oréal is led by Samuel du Retail (General Manager AI, Data and Shared Services); Asmita Dubey (Chief Digital and Marketing Officer); Etienne Bertin (Group Chief Information Officer). Ownership of AI on the org chart: Samuel du Retail, General Manager AI, Data and Shared Services; a new AI transformation team.
- What AI platform does L'Oréal use?
- L'Oréal builds its own AI tools on bought models: L'OréalGPT for employees, CreAItech for marketing content, BETiq for media spend and a Beauty Tech Agentic Platform on Google Cloud. The tools all draw on L'Oréal's Beauty Tech Data Platform (Apr 2026).
- Which AI models does L'Oréal use?
- L'Oréal picks models per tool: Google Gemini, Imagen and Veo in CreAItech, and OpenAI's GPT-4o in Beauty Genius and GPT-Rosalind in research. L'Oréal also uses Anthropic Claude for analytics agents, NVIDIA software and a formulation model built with IBM (2025 to 2026).
- How many L'Oréal employees use AI?
- More than 56,000 of L'Oréal's more than 95,000 employees had L'OréalGPT access in Apr 2026, and about 21,000 used it daily. By then, 73,000 L'Oréal employees had taken the Gen AI for All course.
- What return has L'Oréal reported from AI?
- L'Oréal has reported returns it credits to AI in one function, marketing, which places it at Piloting. CEO Nicolas Hieronimus said on Feb 13, 2026 that the cost of one piece of content made in the CreAItech studio is down 40%. On Feb 20, 2026, Hieronimus said the BETiq media tool generates 10% to 15% productivity on advertising and promotion spend in each country's first year. On rollout, Samuel du Retail, General Manager AI, Data and Shared Services, told the Apr 24, 2026 annual meeting that more than 56,000 L'Oréal employees have access to L'OréalGPT and 73,000 have been trained in generative AI. Du Retail was also reported saying 21,000 employees use L'OréalGPT daily. L'Oréal falls short of Scaling because both returns come from marketing and because it has published no firm-level AI figure. Hieronimus says L'Oréal's AI use cases represent significant productivity on several P&L lines but gives no number. CFO Christophe Babule's expected 20 to 30 basis points of yearly improvement in selling, general and administrative costs is credited to IT transformation and AI together.
Open questions
- Will L'Oréal publish a firm-level AI figure, in euros or margin, at its AI-focused capital markets day? How much of the expected 20 to 30 basis points of yearly improvement in selling, general and administrative costs does L'Oréal credit to AI rather than to IT transformation?
- What does the 40% lower cost per piece of CreAItech content add up to in euros against L'Oréal's €14.2 billion of advertising and promotion, and is the saving kept or reinvested in more content?
- How does L'Oréal measure BETiq's 10% to 15% productivity, and does it hold after the first year in each country?
- What return in cost, headcount or sales do L'Oréal's other deployed tools earn, such as Beauty Genius, the consumer care agents, the shelf-check agents, K-SCAN and the research models?
- Which models run L'OréalGPT, and how is traffic split between Google, OpenAI, Anthropic and other providers?
- To whom do Samuel du Retail, Etienne Bertin and the AI transformation team report, and does an AI office report directly to Nicolas Hieronimus, as one Jun 2026 report said?
- How many AI and machine learning specialists does L'Oréal employ, beyond the 8,000+ digital, tech and data talents it reports?
- Which of L'Oréal's figures for L'OréalGPT use is current, given 56,000+ with access and 21,000 daily in Apr 2026 against 71,000 users in 2025 in the registration document?
- Has the Formulation Foundation Model built with IBM moved from early results into routine use by L'Oréal's researchers, and what has it saved?
Sources
L'Oréal
- 2025 Annual Report: Beauty Tech acceleration with AI
- 2025 Annual Report: Transformation
- 2025 Annual Report: Interview with Nicolas Hieronimus
- 2024 Annual Report: Culture and talents
- CreAItech, L'Oréal's GenAI beauty content platform
- L'Oréal and OpenAI join forces for transformation in beauty with AI (Jun 17, 2026)
- L'Oréal and NVIDIA to accelerate beauty discovery with predictive AI science (Mar 17, 2026)
- L'Oréal and NVIDIA collaborate on next-generation AI (Jun 11, 2025)
- Succession of Jean-Paul Agon as CEO from May 1, 2021
Filings, results and shareholder materials
- 2025 Universal Registration Document: Executive Committee at Mar 1, 2026
- 2025 Universal Registration Document: CEO prospects
- 2025 Universal Registration Document: CEO remuneration criteria
- 2025 Universal Registration Document: ethics risk and Trustworthy AI
- 2025 Universal Registration Document: intellectual property risk
- 2025 Annual Results (Feb 12, 2026)
- 2026 Half-Year Results (Jul 29, 2026)
- Sales at Sep 30, 2026 (Oct 22, 2026)
- Annual General Meeting of Apr 24, 2026
- Letter to shareholders No. 91, AI that serves all (Spring/Summer 2026)
- BoursoLive: L'Oréal and AI, interview with Samuel du Retail
Call and conference transcripts (Quartr via Stock Analysis)
- 2025 results call (Feb 13, 2026)
- CAGNY 2026 conference (Feb 20, 2026)
- dbAccess Global Consumer Conference (Jun 2, 2026)
- H1 2026 results call (Jul 30, 2026)
- Barclays Global Consumer Conference (Sep 8, 2026)
- Q3 2026 sales call (Oct 22, 2026)
- Capital markets day 2025 (Dec 3, 2025)
- H1 2025 results call (Jul 30, 2025)
Partners
- Google Cloud: Introducing Gemini Enterprise Agent Platform, Etienne Bertin quote (Apr 22, 2026)
- Anthropic: L'Oréal advances conversational analytics with Claude
- Microsoft: L'Oréal deploys Beauty Genius with Azure OpenAI Service
- Salesforce: L'Oréal handles support cases up to 64% faster with Agentforce (Oct 20, 2026)
- IBM and L'Oréal to build first AI model for sustainable cosmetics (Jan 16, 2025)
- Galderma: Samuel du Retail biography
News
- Consumer Goods Technology: L'Oréal AI roadmap (Apr 27, 2026)
- Digiday: L'Oréal accelerates generative AI content engine with OpenAI deal (Jun 19, 2026)
- Fortune: L'Oréal's innovation chief says AI rewards companies with history (Jun 22, 2026)
- PYMNTS: L'Oréal CEO touts AI-powered beauty at VivaTech 2025 (Jun 2025)
- Reuters via The Star: L'Oréal to invest $383 million in Indian tech hub (Jan 21, 2026)
- Cosmetics Business: L'Oréal partners with Google Cloud on generative AI content
- diginomica: CES 2024, gen AI and L'Oréal's beauty tech (Jan 2024)
- Finimize: Berenberg sees L'Oréal pulling ahead in AI beauty search (Oct 2026)
- Metis Strategy: interview with Barbara Lavernos
Data confidence
- Confirmed
- Stated in a primary or reputable public source.
- Estimated
- Derived or approximated from public reporting.
- Unconfirmed
- Reported, but not yet corroborated.
L'Oréal profile · Enterprise AI diffusion report by Imagine · facts checked against their sources on Nov 9, 2026
