Aréas Assurances
SQLI designs an AI-powered document chatbot for Aréas Assurances
Aréas Assurances is an independent group of mutual insurance companies operating across the whole of France. It offers tailored solutions for individuals, professionals, and businesses. In certain areas, the company holds a significant body of documentation, with a large volume of procedures and business rules available to its teams.
2024
CONTEXT
Faced with a dense and constantly evolving body of documentation, the Transformation and Information Systems Department (DTSI) of Aréas Assurances sought to improve access to information within the department, particularly for new employees.
CHALLENGES
Several key objectives guided the implementation of the project:
- Reduce the time needed to access information within a complex documentation environment.
- Facilitate the onboarding of new employees through a simple and effective tool.
- Ensure the reliability of the system’s responses.
- Gain autonomy in order to deploy the solution across other documentation domains.
OUR ROLES
To meet these objectives, Aréas Assurances partnered with SQLI to design an intelligent document chatbot, capable of querying a structured corpus and delivering precise, contextualized answers.
The project was structured around the following key stages:
- Functional scoping through collaborative workshops to clearly define the needs.
- Custom development of the chatbot by SQLI’s technical teams.
- Full integration within the existing documentation ecosystem (notably Teams).
- Implementation of a feedback loop to improve the relevance of responses.
- Knowledge transfer to enable Aréas Assurances to autonomously deploy the solution across other domains.
RESULTS
The initial results are very encouraging:
- The chatbot was launched at the end of 2024 and is currently used by around twenty employees.
- A rollout to the entire department is planned by the end of 2025.
- The system now achieves a 95% accuracy rate, thanks to continuous learning based on user feedback.
- Aréas is now autonomous in setting up a new document silo, with a first test scheduled for July 2025.
- The product is configurable and continues to evolve in order to be extended to other teams and use cases.
- The tool is generating growing interest within the organization, particularly for use cases related to support and internal procedures.
The SQLI teams demonstrated strong technical expertise and a pragmatic approach. Despite a limited budget and tight deadlines, we were able to implement an effective and high-performing solution — something that remains rare in AI-related projects, which are often complex, costly, and lengthy to deploy.