Transforming insurance pricing using Amazon SageMaker AI with Eazy Asigurari

Romania’s insurance market is highly regulated, limiting flexibility in how insurers calculate premiums. At the same time, customers can compare prices online with minimal effort. Even small price differences can prompt them to switch providers—particularly in the case of mandatory insurance products, such as motor third-party liability.

As a new entrant to the market, and unlike many established insurers, Eazy Asigurari sought to supplement traditional actuarial inputs with data-driven approaches. This created an additional layer of complexity in a market where pricing needs to carefully balance risk, profitability, and competitiveness.

The company recognized an opportunity to scale its analytical capabilities beyond traditional methods. “Building and scaling our own model infrastructure from scratch would have required significant time and resources,” says Ioana Alexandra Frincu, chief technology officer at Eazy Asigurari.

Using advanced data and ML capabilities on AWS, Eazy Asigurari built its technology infrastructure in the cloud from day one. This foundation helped the company develop a dynamic pricing engine that is powered by data and designed to improve risk accuracy and support long-term growth. “We wanted to embed AI at the core of our organization,” says Ioana Frincu.

Rather than replacing its existing pricing framework, the company is enhancing it with ML capabilities. AWS Glue—a serverless service that makes data integration simpler, faster, and cheaper—handles data ingestion and transformation. To secure data intake workflows, Eazy Asigurari uses Amazon API Gateway—a fully managed service that makes it easy for developers to create, publish, maintain, monitor, and secure APIs.

The company stores its data in Amazon Relational Database Service (Amazon RDS) for PostgreSQL, which provides easy-to-manage relational databases. Meanwhile, Amazon DynamoDB—a serverless, fully managed, distributed NoSQL database—captures logs and model outputs to support traceability, which is an important consideration in a regulated industry.

At the core of the solution is Amazon SageMaker AI, which lets data scientists and developers build, train, and deploy ML models for virtually any use case. To efficiently validate assumptions and test pricing behavior across thousands of customer segments, Eazy Asigurari customized models that can backtest premiums and highlight potential deviations. Then, the company’s actuarial team reviews all ML outputs before informing pricing decisions.

“Advanced pricing models require a large volume and variety of data, exceeding the capabilities of traditional tools,” says Ioana Frincu. “Being able to work with diverse data sources—including synthetic data—and simulate scenarios was a huge outcome for us.”

By using managed AWS services, Eazy Asigurari can focus on refining its pricing strategy rather than managing infrastructure. “Having all our infrastructure on AWS made it easy to modify and optimize components as needed,” says Ioana Frincu. And because AWS has rigorous controls to strengthen data sovereignty and privacy posture, Eazy Asigurari can meet data residency and infrastructure security requirements more efficiently.

Completing a proof of concept within 2 months, the insurer is moving the solution into production. The initial benchmarks are promising: Early model validation confirms alignment with actuarial expectations, and the ML outputs meet the pricing target. This validates that the models can enrich the company’s pricing logic.

Using the new ML-driven workflow, Eazy Asigurari reduced data analysis time by up to 80 percent, accelerating decision-making. The company can test scenarios and benchmark results more consistently, refining pricing strategies faster and more confidently. By helping ensure premiums are competitive, Eazy Asigurari is increasing customer confidence while supporting sustainable, long-term growth.

Aiming to integrate data-driven decision-making more deeply into its underwriting operations, the company will expand its use of advanced analytics across additional insurance products. “This type of project aligns with our core philosophy to not be AI enhanced, but to be AI embedded,” says Ioana Frincu.