CLUB MED & THIGA

Improving delivery efficiency for product teams using a GenAI Agent

GenAI for Teams
Operational Excellence
Hospitality
Club Med 3

Context

AI is profoundly transforming the market and redefining performance standards. True to its pioneering DNA, Club Med has chosen to fully embrace this shift by laying, with Thiga, the first building blocks of an AI-ready organization.

The ambition is clear: empower teams to reinvent the way they work by leveraging AI agents capable of automating repetitive tasks, freeing up time for innovation, creativity, and operational excellence.

In this joint mission, the agents were built by Thiga in close collaboration with Club Med’s teams, ensuring a perfect fit with both business and technical needs. Beyond the technological dimension, Thiga also supports change management: adoption of new practices, upskilling of employees, and seamless integration into daily operations.

This approach has made it possible to identify and test concrete, high-impact use cases across the entire product and tech delivery chain, from creating User Stories to generating release notes, understanding APIs, and automating tests. By capitalizing on solid governance and smooth integration with existing tools, Club Med has taken a structural step forward: experimenting, measuring, and optimizing, to lay the foundations of an AI-augmented organization, where technology becomes a true driver of collective and sustainable performance.

Challenges

  • Identify the most value-creating use cases: rapidly pinpointing high-impact AI opportunities, replicable across the organization and aligned with business priorities.
  • Design a robust and integrated agentic architecture: ensuring compatibility with Club Med’s existing technology stack, while guaranteeing security and scalability.
  • Engage product teams in co-construction: involving team members from the design stage, fostering ownership of new practices, and supporting change management for sustainable adoption.
  • Measure and demonstrate tangible benefits: proving the effectiveness of AI agents through precise KPIs (productivity, quality, support reduction) and validating the added value of each experiment.

Our approach

  • Mapping product & tech processes to identify AI opportunities

    • Internal observation and analysis: interviews with teams, observation of rituals and interactions to detect pain points
    • Mapping of existing processes and assessment of their complexity/recurrence to identify automation opportunities
    • Technical interviews and alignment with IT to define prerequisites and constraints for experimentation
  • Use case prioritization: user stories, APIs, release notes, etc.

    • Identification and selection of relevant use cases via prioritization workshops (e.g., US, API, release notes, code)
    • Detailed formalization of selected use cases with benefit hypotheses and measurement KPIs
    • Evaluation of feasibility, impact, and scalability of candidate use cases
  • Building and deploying first agents

    • Setting up the agentic ecosystem: environments, configuration, and integrations (N8N, databases, internal tools)
    • Development of agents: flow creation, source integration (Jira, GitHub, APIs, etc.), and prompt engineering
    • Testing with business and technical teams to optimize agent performance
  • Red 25

    Iterative sprints with human-in-the-loop feedback

    • Defining scope and objectives for each sprint (build, optimize, measure)
    • Experimentation with early adopters and continuous feedback collection to improve the agent
    • KPI analysis at the end of each sprint and incremental adjustments to agent workflows
  • Red 21

    Measuring of impact and continuous optimization

    • Systematic KPI tracking (productivity, quality of deliverables, support reduction, team velocity)
    • Synthesis and sharing of results with stakeholders (impact, limitations, improvement areas)
    • Gradual integration of new sources or variants to enhance agent performance
  • Red 16

    Laying the foundation for multi-agent governance and scaling

    • Defining technical, legal, and organizational prerequisites for industrialization
    • Establishing a Product Operating Model and shared governance to supervise AI agents
    • Planning large-scale deployment (candidate teams, roadmap, handover to internal teams)

Our impact

  • Time savings on content creation and support tasks : From 3-5 days to 0,5 day
  • Better documentation quality and usability
  • Faster time to market
  • Adoption of a scalable framework for AI agent deployment - in progress

Would you like to know more about this project or another?