AI + Data Is a Force Multiplier in P&C

The power of machine learning is amplified by the growing market of third-party data available to train and refresh models.

An artist’s illustration of artificial intelligence (AI). This image depicts how AI could be used in the field of sustainability from biodiversity to climate.

Data-driven decision-making has long been the goal of P&C commercial lines carriers. There has never been a shortage of data within a carrier’s own walls, and decades ago some sought to create a competitive advantage through the use of predictive models. However, it was a considerable challenge to amass and normalize enough structured data to train a model. The handoffs to run and use a model were manual. And keeping the model current—by retraining it with newer data—ran into the same challenges.

Despite these early hurdles, carriers saw the value of using models to evaluate risks and identify the new business submissions in the queue with a higher probability of winning. Risk analysis models eliminated discretion in comparing exposures with target account guidelines. Predictive reserving models avoided being solely reliant on each claims adjuster’s experience to recognize the losses that looked simple at intake but carried all the hallmarks of a complex and costly claim.

See also: Why AI Is a Game Changer

An Inside Look at AI in Commercial Lines Carriers

Resource Pro Insights’ newly released research, “Artificial Intelligence in P&C Commercial Lines: Carrier Plans, Perceptions and Potential for High-Value Use Cases,” offers a comprehensive look at AI within commercial lines carriers today.

We include robotics process automation (RPA) in our AI research. While not everyone considers this an AI technology, RPA has proven itself to be an effective, adjacent-to-AI solution for carriers to automate repetitive actions. Our research reveals that RPA is well-embedded within commercial lines, with most carriers being in the investment phase of planning, piloting or running in production.

AI solutions are helping commercial lines carriers realize new value across the insurance lifecycle, with even more potential in the future. For example, conversational AI allows insurance carriers to offer self-service and personalized product education, enhancing customer experience. The value of RPA for billing is on the rise, to achieve both precision and speed in producing invoices and booking receivables, functions that can span multiple, disconnected systems. Advances in voice systems that include multilingual natural language processing are removing friction in policy servicing interactions. Image recognition and computer vision are giving carriers the ability to expand the scope and geographic reach of their loss prevention services.  

Machine learning in the insurance industry plays a key role in helping carriers make data-driven decisions. More than 75% of carriers have plans or pilots or will be using it in production this year. The power of machine learning is amplified by the growing market of third-party data available to train and refresh models. Within underwriting and risk management, these third-party sources enable carriers to automatically augment risk profiles and verify submitted data -- for example, SIC/NAICS codes. Machine learning can score each risk, apply more refined straight-through-processing rules and triage underwriting referrals based on a model prediction of those most likely to bind.

The value carriers believe AI can offer in the commercial lines submission process is noticeably higher this year than last. New business applications and policy change requests contain unstructured and structured data in a seemingly infinite array of formats. Machine learning is able to normalize submission data and validate or augment using third-party sources. What could have taken days can now occur in minutes.

Commercial carriers also see value in using AI for predictive claim reserving, an area claims organizations have been modeling for years. The earliest efforts by carriers had their data scientists creating models that were manually run after each new claim was registered. Now there are solutions with models based on both internal and external data. Some not only automate predictive reserving for each new claim but also run throughout the life of a claim, triggered by changes to the loss information.

The expected value of AI for commercial claims fraud monitoring and detection is lower than last year, but still high overall. Carriers appreciate that AI can automate a more thorough approach to continuously monitor open claim files for potential fraud. Many available external data sources play an essential role in offering carriers a nationwide lens on the bad actors and other warning signs.

Learn more about the current state of AI for property and casualty commercial lines carriers by reading our new research report, “Artificial Intelligence in P&C Commercial Lines: Carrier Plans, Perceptions and Potential for High-Value Use Cases.” 

Meredith Barnes-Cook

Profile picture for user MeredithBarnes-Cook

Meredith Barnes-Cook

Meredith Barnes-Cook is a partner at ReSource Pro Consulting.

She leads a growing consulting practice with a focus on carrier advisory services, leveraging decades of industry knowledge, digital expertise, change management and entrepreneurial spirit to help insurers navigate the ever-evolving landscape of the insurance industry.


Read More