Artificial intelligence (AI) capabilities are rapidly evolving. There are many different types of AI applications for insurance, but for improving underwriting capabilities and controlling loss ratios, the use of predictive and prescriptive analytics could be especially rewarding. Early adopters stand to gain a significant competitive advantage.
Investopedia1 defines AI as the simulation of human intelligence by software-coded heuristics. Machine learning is a subset of AI involving computer programs that learn and adapt without human assistance.
AI has earned mainstream popularity in recent months thanks to tools like ChatGPT. However, AI is not exactly new. According to Britannica2, AI began in the mid-20th century with the work of Alan Turing, who predicted computers would eventually become good at chess. In 1997, his prediction came true when the Deep Blue computer built by IBM beat the reigning world chess champion. Since then, AI has come a long way.
Investopedia3 defines data analytics as the science of analyzing raw data to draw conclusions about the data. There are four main types of analytics:
All four types of analytics can be useful in insurance. For example, descriptive analytics can tell us whether claims have increased, diagnostic analytics can tell us whether changing weather patterns are responsible for an increase in claims, predictive analytics can tell us whether an increase in claims is likely in the next year, and prescriptive analytics can tell us what we should do to minimize claims.
Insurance is all about making high-stakes decisions based on what you think might happen in the future, and with predictive analytics, insurers have been able to center these decisions on hard data. But according to Graham Bartholomew, Guidewire’s Director of Pricing Solutions, an analytics evolution is occurring at best-in-class companies where the focus is now turning from predictive to prescriptive analytics, which recommends prescribed actions along with the potential outcomes.4
Property Casualty 3605 uses an example involving water loss. Prescriptive analytics can predict the likelihood of water loss and recommend where water sensors should go.
In another article, Property Casualty 3606 explains that analytics can also help with policyholder retention. It uses the example of a company learning that 85% of policy lapses come from 10% of customers. The company can then use this information to improve retention.
And in looking at the example of a car accident, prescriptive analytics would go beyond just estimating the total incurred loss to actually detailing what repairs would need to be made and recommending an auto repair shop.7
We appear to be at a turning point in AI development. In LLMS: What to Do Next, Celent8 predicts that approximately 20% of companies will be early adopters, embracing new AI technology like ChatGPT by 2024. By 2026, up to half of companies will have embraced this technology.
Insurance companies that don’t act quickly may find themselves at a disadvantage in the near future. This is true of both large language models like ChatGPT and advanced predictive analytics tools.
Knowledge is power, but, in the case of predictive analytics, knowledge could also mean profit. In an interview with Business Insurance9, Stan Smith, CEO of Gradient AI, discusses how insurers can use AI to reject bad business that competitors might pick up if not using AI. If the insurers using predictive analytics are able to make better decisions than other insurers, the difference in combined ratios could quickly become significant.
Insurers could benefit from this break right now. According to the Insurance Information Institute10, the property and casualty insurance sector is expected to report a combined ratio of 105.6 for 2022, indicating an underwriting loss. AI-powered analytics could help insurers return to underwriting profitability faster.
Retention is another pressing issue. According to the Independent Insurance Agents of Dallas11, the insurance industry has a retention rate of 84%. For top companies, the retention rate is usually around 94% to 95%. Since attracting new customers is time consuming and expensive, policyholder churn is a major drain on resources. If insurers can boost retention using analytics, they can also protect their bottom line.
Per Reuters Insight’s Technology Benchmark Survey, insurer investment in advanced technologies is ramping up, with Analytics & AI ranked highest as the technology type that delivered the best ROI and was expected to receive heavy investment over the next two years.12
AI progress is occurring incredibly fast – so fast that the Future of Life Institute13 has issued an open letter calling for a six-month pause on AI developments. Elon Musk and Steve Wozniak are among the many people who have signed this letter.
Companies may not pause AI development voluntarily, but stricter regulations could be coming. According to Digital Insurance14, the Colorado Privacy Act establishes guidelines for AI usage. Although it applies primarily to life insurance, the Colorado Department of Insurance has said property and casualty insurers should expect similar rules. Other states may follow suit. Meanwhile, the NAIC is working on its own guidelines for AI.
This is an exciting time for AI development. It’s also an exciting time for digital payments. One Inc is the leading digital payments network for the insurance industry, providing seamless digital payment experiences that allow insurers to deliver more to their customers.