Turkish insurer Neova Sigorta, in collaboration with data and AI company SAS and SAS Partner Sade Software & Consultancy, has launched an initiative to offer better auto insurance premium prices to up to 95% of its customers.
The project, among the first of its kind in the Turkish market, will use artificial intelligence in the form of advanced machine learning (ML) to optimise how it prices auto insurance.
This advancement is projected to save Neova Sigorta’s customers money – and decrease overhead costs for the insurer.
Neova Sigorta selected SAS Dynamic Actuarial Modeling, a pricing solution with AI-based premium modeling for general and life insurers, as its platform of choice for its transformation. Software and consultancy firm Sade Yazılım will be the initiative’s integrator.
Similar collaborations between SAS and SADE demonstrated that the right pricing policy can increase sales up to 15% and decrease the insurer’s combined ratio by 10%.
The deployment, expected to last between six and eight months, is intended to help Neova Sigorta offer more appealing prices to new and existing customers. It is also expected to significantly increase market share in new regions and bolster renewal rates in established regions.
“In our business, we always prioritize customer comfort and satisfaction above all else, and we know that affordable premiums are one of the main factors that contribute to customer happiness and retention,” said Neslihan Neciboğlu, Neova Sigorta CEO and board member.
The CEO said that among the several platforms they evaluated for the implementation of this project, SAS Dynamic Actuarial Modeling was the most comprehensive platform as it enabled them to train, deploy and automatically monitor the performance of high-accuracy machine learning models.
“We also appreciated that SAS’ solution capabilities can be extended to other critical functions such as next best offer generation and insurance fraud detection, aiming to improve customer experience and reduce prices even further,” she added.
Stu Bradley, senior VP of risk, fraud and compliance solutions at SAS, said that fair, accurate and timely insurance policy pricing is a necessity for insurers in all markets.
“Consumers often look for more affordable insurance options amid inflation and other economic turbulence,” said Bradley. “If insurers can’t offer competitive pricing, they risk losing customers to the competition – and agents often follow when they can’t achieve their goals. This can severely damage an insurer in mere months.”
For example, auto insurance pricing has historically been based on generalized linear models (GLM). Although these models have good interpretability – performance is predictable and easily explained – the accuracy of estimates tends to be limited, ultimately resulting in higher prices and lower sales.
In addition, limited sample sizes and risk assessment policies significantly deteriorate the quality of the data used to train GLMs, which can cause errors and introduce bias.
Looking to modernize, Neova Sigorta will switch to an alternative approach based on sophisticated ML algorithms to price its auto offerings.
Unlike GLMs, machine learning algorithms do not make any assumptions about the properties of the data the models are using. These algorithms also consider more variables that define customer behavior and are designed to extract more granular patterns in the data. Therefore, ML algorithms produce much more accurate – and realistic – results.
