Matmut relies on generative AI to meet real needs, not trend
An AI reserved for employees, not customers
For ethical reasons and to avoid risks such as AI hallucinations, Matmut refuses to place generative AI directly in the hands of its customers. Instead, the insurer prefers to leverage this technology to support its employees by developing intelligent agents that will assist them in their daily tasks starting in 2025–2026.
A discreet yet active presence in the AI ecosystem
On March 24, during the Hub France IA plenary by Gaia-X in Bercy, the financial sector — including banking and insurance — remained low profile. Yet Matmut is making progress. In February, Olivier Monnier, Chief Data & AI Officer, acknowledged at the Data Mobility Day: “Sharing is not in our DNA,” highlighting the sector’s distinctive culture around data.
A cultural shift: from business wants to actual needs
For the past four years, Matmut has structured its data governance around a Data Office that operates independently from the IT department. But according to Olivier Monnier, the real transformation has been the shift from experimentation to concrete production, with a strict filtering of business projects.
Each request must now meet three criteria:
Why is the project necessary?
What is the risk of not doing it?
What value will it generate?
This framework has led to an 80% reduction in the number of requests — while increasing deliverables by 30%, demonstrating the effectiveness of this more focused approach.
A value chain–driven Data organization
To address business needs, Matmut has structured its Data Office around several complementary units:
A department that works directly with business teams to translate their needs into data terms
Data Engineering teams
Experts in Data Science and AI
This cross-functional organization enables a fast transition from business requirements to production deployment.
Raising executive awareness to structure the use of GenAI
As early as 2023, Olivier Monnier alerted the Executive Committee (Comex) to the challenges of generative AI: regulation, integration into information systems, and complementarity with other AI technologies. He emphasizes, “We’re not going to use GenAI when Machine Learning is enough.”
AI agents to automate advisors’ tasks
Matmut is focusing its efforts on concrete use cases where generative AI proves to be both more effective and cost-efficient. One of its flagship projects is the development of AI agents to assist advisors by automating tasks such as note-taking, quote creation, and contract drafting.
The goal is to free up time so that advisors can fully focus on customer relationships. These agents are expected to be operational by 2025–2026, within an ethical framework that also takes environmental impact into account.
source of the news : ZD NET
