Matmut Derives Data Insights Three Times Faster

Having a single platform in DataRobot is a revolution for us. Everything is just simpler.

To elevate nearly every area of the company, the Matmut Group relies on data. But the major French insurance company must derive insights within the limits of stringent privacy regulations.

In recent years, Matmut’s data lab built predictive models with a single Jupyter notebook. Every step of the process was manual and required considerable coding. The company sought a single solution to reduce the effort and enable collaboration between data scientists and the business.

“Placing data at the center of everything is the key to success at Matmut, but it was slow and complicated without the right tools,” said Olivier Monnier, Chief Data Officer. “We needed to be more responsive and proactive to the needs of the business.”

Automating Predictive Analytics End to End

The Matmut Group deployed DataRobot AI Platform to automate the end-to-end lifecycle, from data preparation to modeling to monitoring.

Beginning with a proof of value (POV), the platform delivered efficiency gains across every step of the process. On the front end, Feature Discovery automatically generated new features based on the dataset, expediting data preparation.

Likewise, the platform automated modeling with machine learning, along with deployment and updates. The combined time savings across the lifecycle amounted to a three times productivity gain over the team’s previously manual approaches.

DataRobot University: Key to Success

From the start, Matmut embraced DataRobot University to increase users’ proficiency on the platform. On average, each team member has taken 10 courses covering various aspects of the solution.

“We needed to ensure that everyone in the company knows how to derive value from the tool and use it efficiently,” Monnier said. “For that, DataRobot University is the key to success for Matmut.”

Plus, the team at DataRobot guided Matmut every step of the way.

DataRobot is not a solution provider. They are a partner. It’s very important. It means they were present prior to the contract, during the contract, and after the contract. That’s impressive.

With automation and ease of use for even non-data scientists, DataRobot AI Platform has effectively changed the way that Matmut works. Data scientists, machine learning engineers, business analysts, and actuaries all collaborate on business challenges.

“We are creating an ecosystem to ensure we have something robust between the business and IT,” Monnier said. “And that’s building a stronger relationship between our team and the business units.”

Now, Matmut is applying AI to numerous areas of its business:

• Forecasting customers’ needs for auto loans
• Predicting the volume of claims to optimize staffing and maintain high customer service levels
• Understanding natural disasters for business planning and pricing

‘Everything is Simpler’

As the team provides insight to the business, business units increasingly look to the team for trusted answers. With that demand, Matmut has set an ambitious goal of applying the AI platform to more than 60 use cases — and with fewer resources than it would need otherwise.

“Without DataRobot AI Platform, I would need to recruit more data scientists and more data engineers,” Monnier said. “Having a single platform is a revolution for us. Everything is just simpler.”