This interview is part of our new AI in Insurance series, where we interview the world's top thought leaders on the front lines of the intersections between AI and insurance.
In this interview, we speak with Mike Finegold, Chief Data Scientist at Fulcrum Analytics, to understand how his company is using AI to transform insurance, and what the future of the insurance industry holds.
1. What's the story behind Fulcrum Analytics? Why and how did you begin?
MF: For over 25 years, Fulcrum Analytics has been providing advanced data science consulting that includes building sophisticated custom apps for our clients. From the beginning, we have taken a pragmatic approach that focuses on flexibility, custom-tailored solutions, and practical data science resulting in actionable insights and applications that improve, not replace, large investments.
2. Please describe your use case and how your company uses artificial intelligence:
MF: A common problem for property and casualty insurers is streamlining their insurance policy renewal process. This process involves collaboration between price modelers, actuaries, and underwriters and has historically been inefficient and prone to human error. In some cases, the process involves so much manual labor that there aren’t enough resources for proper data audits and model updates. On top of that, it’s very difficult for some carriers to monitor the effectiveness of pricing models, let alone the impact of an underwriter’s decision to renew at rates different from what the models suggest.
A complete solution to this problem requires more than just AI: visualizations that allow underwriters to dive into accounts to understand what led to pricing recommendations; scenario planning tools so underwriters can see the impact of different rates on account P&L; benchmarking against similar policies; a common interface for pricing, actuarial, and underwriting; and more. But AI plays a critical role in at least two areas: automating the data curation process and monitoring the performance of models and underwriter decisions.
For pricing models to be effective, first and foremost the data needs to be reliable. Manually checking and validating every variable to ensure that it rolls up properly to the relevant modeling unit has historically been very labor-intensive and left countless errors undetected. Moreover, models must be adapted to changes in the underlying insured population. We use AI to automatically detect data anomalies and monitor changes in the underlying population. In using AI, data issues can be detected and corrected before the model results are reviewed. As a result, pricing teams can devote resources to building new models – not on an arbitrary schedule – but exactly when a change in underlying data necessitates it, improving both process efficiency and model accuracy.
On the other end of the process, a proper renewal system should capture the output of every pricing model, actuarial adjustment, underwriter decision and ultimately the profitability of every policy. We use AI to pinpoint which parts of the process need to be tweaked to optimize performance. Are the pricing models ineffective for certain LOBs or customer types? When underwriters override a recommendation not to renew, does this lead to unprofitable accounts? Do the underwriters’ account knowledge and the statistical models complement each other to ultimately lead to the right decisions? These are all questions that can be answered through the deployment of AI.
We developed our ClearGrade tool to dramatically reduce inefficiencies and improve the effectiveness of the entire renewal process, and AI plays an ever-increasing role in our solutions.
3. Could you share a specific customer/user that benefits from what you offer? What has Fulcrum Analytics done for them?
MF: Recently, we helped a leading global P&C carrier revamp their renewal process using ClearGrade, resulting in a dramatic decrease in the turn-around time for pricing policy renewals. The speed to price the renewals improved due to the shared interface, reduced errors, greater transparency in pricing inputs, and quicker comparison to like policies that ClearGrade provides the client.
Moreover, the models are much more reliable due to automated data curation that catches more data errors and anomalies. As a result, the carrier is making smarter decisions about their renewals and generating revenue dollars more efficiently.