How Spain’s leading insurer scales AI across the business while meeting governance demands
VidaCaixa turns to DataRobot to accelerate modeling for high-value use cases and ensure AI governance.
We’re faster with DataRobot. Every four models developed with DataRobot equals one extra data scientist we would otherwise need.
AI in insurance: balancing risk and return
Leading insurer VidaCaixa is embracing AI to make better business decisions from underwriting to payouts to call center routing. With more than 100 years in business, VidaCaixa leads the market as Spain’s largest insurance company, serving 6.8 million customers with life insurance and pension plans.
AI plays a critical role in ensuring that VidaCaixa can operate more efficiently and continue providing their customers with the best possible experience. For Data Analytics Director and Chief Data Officer Jordi Vidal Araujo, increasing demand for AI use cases requires balancing risk, regulatory requirements, and company standards.
“AI governance is more important than ever before,” he explained.
Building a scalable, effective AI strategy demanded tooling that streamlined how VidaCaixa built and governed models.
DataRobot accelerates the entire modeling process, allowing us to evaluate many models and outputs and find the best ones in a few minutes instead of weeks. The platform also helps ensure that models meet our AI governance standards.
Scaling AI in the insurance industry
VidaCaixa chose DataRobot to enable the team to keep pace with demand and ensure governance in the tightly regulated insurance industry.
“DataRobot accelerates the entire modeling process, allowing us to evaluate many models and outputs and find the best ones in a few minutes instead of weeks,” Araujo said. “The platform also helps ensure that models meet our AI governance standards.”
The ease of creating models empowers the team to try more variations at a faster pace. Rather than building one or two models to solve a business problem, they can iterate with multiple options quickly for champion-challenger testing.
Via the DataRobot Python API, they bring numerous features into every model and can evaluate and weigh the influence of each. Araujo appreciates that they can check the data behind each model within DataRobot without having to switch tools.
“We’re faster with DataRobot,” Araujo said. “Every four models developed with DataRobot equals one extra data scientist we would otherwise need.”
They can save models as blueprints in the AI Catalog, preserving the steps, techniques, and algorithms behind each. They can then reuse or repurpose them for the same or other models.
Araujo notes that junior data scientists can develop models easily, helping them tackle the growing demand from the business.
“It’s easy to understand how DataRobot works,” Araujo said. “It’s changing our mindset. You don’t need a lot of information about how models work or need to know how to code.”
As VidaCaixa scaled predictive and generative AI, the company kicked off a strategic initiative to define the standards and policies around AI use. Behind all DataRobot models, they have documentation for compliance reporting and internal governance, and built-in bias mitigation and fairness guardrails so models remain balanced.
Our experience with DataRobot data scientists has been amazing. When we’re starting new models, they provide a lot of knowledge and support on how to approach projects. If you want to get faster, and improve your KPIs and trust in your models, DataRobot is a good partner.
Saving 30 minutes per claim
The analytics team supports diverse AI use cases that bring value throughout the organization:
Underwriting – Underwriting models help VidaCaixa find the optimal balance of risk and return in extending policies to customers. With feature engineering in DataRobot, they weigh numerous factors for faster and more effective decision making that reduces costs considerably.
Insurance payments – Modeling guides VidaCaixa in assessing life insurance payouts. The company can now automatically make payout decisions in one-third of cases. By replacing manual review, VidaCaixa saves approximately 30 minutes per claim and is still confident in the decisions guided by DataRobot, Araujo said.
As the company explores new use cases, they value the knowledge and support of data scientists at DataRobot.
“Our experience with DataRobot data scientists has been amazing,” Araujo said. “When we’re starting new models, they provide a lot of knowledge and support on how to approach projects. If you want to get faster, and improve your KPIs and trust in your models, DataRobot is a good partner.”
