Article img

Radar 4.4: New version of Willis Towers Watson’s pricing software delivers enhanced governance and security controls

  • Published by:

  • Topics:
    • InsurTech
    • Launches/New Products
    • Rates
    • Risk Modelling

Willis Towers Watson has released an updated version of its Radar pricing software...

Radar 4.4 introduces new security features that provide a more sophisticated governance process, as well as improvements to the software’s machine learning and optimisation capabilities.

 

"Insurers are under constant pressure to deliver leading-edge pricing techniques just to maintain market position, yet ensuring appropriate governance remains critical,” said Colin Towers, Product Leader for Radar at Willis Towers Watson. “This latest release provides significant analytical improvements, as well as the governance and security advantages crucial to operational efficiency, helping insurers to deliver improved speed to market and pricing accuracy.

 

“The risks inherent to a business adopting open source software for its key capabilities can be significant,” added Towers. “In addition to potential security risks, a company is also vulnerable to solutions which lack transparency and are subject to arbitrary change, which can have a significant impact on the ability to transact business. Sophisticated governance of these processes has to be part of the DNA of any business that wants to be resilient and successful in the digital world.”

 

Radar 4.4 implements new fine-grained security controls that simplify the application of the pricing software, particularly when used in larger teams and organisations. Group access controls allow, for example, analytical and pricing teams to ensure appropriate segregation of work while ensuring the entire business can see the full picture. When combined with Radar Live, it ensures an organisation can safely deploy innovative insights from machine learning models to market in minutes.

 

Other updated features of Radar include enhancements to the Elastic Net method, a machine learning tool, to allow hyperparameter searching. This simplifies the modelling process by making it quicker and easier for users to assess what parameters to include within a traditional rating tariff, as well as reducing modelling sensitivity and risk, freeing up valuable resources to concentrate on making quick and informed pricing decisions. Radar’s ratebook optimisation has also been upgraded, resulting in substantial speed improvement. For the first time, this release also includes the option to switch the application into Chinese.