A new conversational service on WhatsApp, powered by generative artificial intelligence, successfully launched in Brazil and recently rolled out in France, along with a Hackathon to support its teams.
True to its pioneering spirit, and in line with one of its five strategic pillars: “Happy Digital”, Club Med implemented a data strategy in 2022 with the deployment of its own Data Factory, a data management service platform. Thereafter, in March 2023, it committed to the development and implementation of a generative artificial intelligence strategy to serve its clients and employees.
The first use case of this generative AI strategy was intended to support the Club Med agency teams, called “Travel Experience Designers” (TEDs), in their daily tasks, by simplifying access to information as they deal directly with customers. The pilot of this project was launched in September 2023, and since then, new use cases have been deployed, including:
- A translation engine in supply tools and in the product information database.
- The automatic indexing of images in the internal media library.
- The semi-automated programming to transform legacy code into modern code.
Club Med’s strategy regarding artificial intelligence is deployed aligning with an ethical approach as Club Med have set up an ethics committee chaired by Professor Jean- Gabriel Ganascia, a computer scientist and philosopher specializing in artificial intelligence and a professor at the Paris Sorbonne University.
WhatsApp and generative AI to enhance the customer experience
One of Club Med’s new objectives with generative AI is to support and enhance the customer experience by making conversational tools (WhatsApp, Messenger, We Chat, etc.) a new booking channel, offering an instant and personalized service at any time. This allowsTravel Experience Designers (TEDs) to focus on higher-value tasks by delegating the most basic ones to AI. Thanks to this new conversational tool, customers can ask questions via WhatsApp, request quotes for their future stays and receive real- time answers within seconds at any time of day or night.
“This new tool, named GM Copilot, relies on LLM (Large Language Model) technology, specifically Microsoft Azure OpenAI, and a technology known as RAG (Retrieval Augmented Generation), both directly integrated into WhatsApp. We are proud to be among the pioneers in the application of this technology,” said Quentin Briard, CEO Marketing Digital G Technology at Club Med.
This initiative was first launched in Brazil in early 2024, before being deployed in Singapore, Malaysia, the United Kingdom, Belgium and Switzerland. More than 30,000 conversations have already taken place via GM Copilot with promising results.
Given this success, GM Copilot was launched in France on 19 November 2024.
The deployment of the chatbot is expected to be rolled out to all markets in 2025.
A Hackathon to leverage AI for improving team efficiency
As part of its ongoing strategy of generative AI development, Club Med organized a three-day Hackathon in June 2024, in partnership with Microsoft.
“There were multiple goals for this hackathon: to familiarize technical and business teams [sales, marketing, HR, legal, etc.] with AI-related challenges; to foster collaboration between different teams and profiles on a common project; and to develop projects addressing specific needs and providing solutions to irritants,” said Siddhartha Chatterjee, Global Chief Data Officer at Club Med.
Several months were needed to prepare for this hackathon, from sourcing the different use cases to identifying and preparing the necessary data.
Six projects, involving six different business teams, were studied and their relevance and feasibility confirmed by the teams involved: legal experts, data scientists, architects, data engineers, product owners, and Microsoft coaches.
- Sales teams: The hackathon focused on data sourcing and the vectorization of databases to power the internal chatbot used by the sales teams in their product information searches.
- Purchasing: The project aims to improve operations management by facilitating the collection of product pricing data, along with tracking their fluctuations, to assist buyers in their procurement decisions.
- Human Resources: A prototype was developed to simplify the assignment of staff to different Resorts each season.
- Marketing: A writing assistant was created and trained in the brand’s terminology and tone of voice to help teams in content creation.
- Legal: An internal RAG model was developed to simplify legal document research.
- IT: Models are being created to assign more efficiently IT support tickets.
Three of these six projects will be launched in production in the coming months. The writing assistant is already being used by French-speaking teams with satisfactory results and will soon be deployed in other languages. The projects related to purchasing, staff assignments, and the RAG model, which require further developments, are expected to scale up in 2025.