L'Oréal O+O case study
4 min read
How L'Oréal O+O accelerated the adoption of Claude Code and Claude Cowork with a three-day hackathon
Claude Code and Claude Cowork at the L'Oréal O+O build hackathon, 8 to 10 June 2026
.png)
150 tech and product experts trained, 12 prototypes built and more than 200 active Claude Code and Claude Cowork users four months later. A look back at a hackathon designed and run by AI Partners to turn AI experimentation into hands-on skills and new business uses.
200+
active Claude Code and Claude Cowork users across O+O teams
−50%
time spent on some everyday tasks, as estimated by L'Oréal's teams
150
tech and product experts trained in three days
Key facts
Industry: Beauty and cosmetics, L'Oréal O+O (the tech and product teams behind product and commerce applications).
The challenge: After a first edition focused on exploring generative AI in 2025, L'Oréal O+O wanted to take the next step: enable its teams to design AI agents and bring these technologies into their day-to-day work.
The solution: AI Partners designed and ran a three-day hackathon, from 8 to 10 June 2026, bringing 150 tech and product experts together around Claude Code and Claude Cowork. Working in cross-functional teams, participants tackled three concrete business use cases, with technical and methodological support throughout the event.
Adoption: Four months later, more than 200 O+O employees use Claude Code and Claude Cowork in their day-to-day work. Dozens of skills built in-house are now used by the teams.
Key results:
- More than 200 active users of Claude Code and Claude Cowork.
- Dozens of skills built and used every day.
- Up to 50% time saved on some tasks, as estimated by L'Oréal's teams.
- 12 working prototypes built in three days.
- Participant satisfaction of 9.5/10.
The challenge
Moving from experimentation to building
In 2025, a first edition of the O+O hackathon let L'Oréal's teams discover what generative AI could do and try out new tools.
In 2026, the ambition was to go further: from discovering the technology to mastering it in practice.
O+O teams build and evolve product and commerce applications used across the group. To bring agentic AI into their work, interacting with a model was no longer enough. They had to learn to design agents, connect them to company data and get product and technical teams working together effectively.
The challenge was twofold: build in-house skills and turn identified use cases into concrete solutions.
The goal: enable employees to build with AI themselves and make these new tools their own for the long term.
The solution
Three days to turn use cases into working prototypes
AI Partners designed a hackathon centred on practice, learning and collaboration between tech and product teams.
From 8 to 10 June 2026, 150 L'Oréal O+O employees came together around three business use cases, identified and prepared upfront with L'Oréal's teams: Cat Agentic, Field IQ and PM Experience.
The required datasets were prepared before the event so participants could work on real problems from day one.
Two principles shaped the hackathon.
1
A shared technology environment built on Claude
Each participant was trained on the tool best suited to their role:
- Claude Cowork for product managers: framing needs, writing specifications and prototyping user journeys.
- Claude Code for developers: designing agents, building skills and connecting solutions to data through the Model Context Protocol (MCP).
Employees learned on the tools they now use in their work.
2
Mixed teams to speed up collaboration
The 150 participants were split into 12 squads combining technical and product profiles.
For three days, AI Partners experts supported them at every step: scoping needs, designing solutions, solving technical issues and reviewing architectures.
The aim was also to pass on a working method, so teams can design and evolve their own AI solutions.
At the end of the hackathon, each squad presented its prototype to a jury of O+O leadership team members.
The results
12 prototypes in three days, and adoption that continues every day
In three days, the 12 squads built 12 working prototypes addressing the three priority use cases identified with L'Oréal. Participants rated the event 9.5/10.
Four months later, more than 200 O+O employees use Claude Code and Claude Cowork in their daily work. The teams have also built dozens of skills that they use regularly.
According to estimates from L'Oréal's teams, these tools cut the time spent on some tasks by 50%.
The hackathon also changed ways of working. By bringing developers and product managers together on shared projects, it encouraged more direct collaboration and a better understanding of what agentic AI can do.
For another deployment example, read our client case: How Trench Group deployed 3 Claude-powered applications in production.
What they say
"It helped us reach a rare level of maturity in using these tools well. I can only recommend this experience to everyone."
Romain Dardour, Head of Product, L'Oréal Beauty Tech
"It was pure building. We had them work on real projects, we prepared the datasets: we really built."
Thomas Spitz, Co-founder and CEO, AI Partners
AI Partners offering involved
AI Academy
AI Partners helps organisations upskill their teams and put artificial intelligence to work in their day-to-day operations.
For this hackathon, AI Partners was involved from programme design through to participant support, combining hands-on training, technical expertise and work on business use cases. Employees came away with the skills to use these tools on their own.
