AI & Machine learning in P&C insurance: Technology, use cases, and opportunities

It's no secret that success in the Property & Casualty (P&C) segment relies on the accuracy of the underwriting process, and how efficiently claims are collected. And while this sounds simple on the surface, the number of factors affecting underwriting is becoming more complex. On the flip side, there's more data available about insurance customers than ever before. P&C firms are at the cusp of an incredible opportunity: by leveraging customer data and applying a layer of cognitive intelligence, they can all but remove the unpredictability that traditionally marks underwriting. This will help to get ahead of competitors and boost revenues.

And glancing at the market as it stands today, the timing couldn't be better. There is a massive difference between the top quintile and the second quintile when it comes to Returns on Surplus (RoS). While 11.6% is the average level of returns for the industry, the second quintile comes in at a modest 11.3%, while at the top, we have a spectacular average of 30.2%.

Clearly, when it comes to the P&C sector, there's no place to be but at the top. The difference between loss ratios for leading and laggard performers comprises a sizeable 28 percentage points. By focusing on underwriting accuracy and claims collection efficiency, firms can boost their operating performance and thereby climb up the ladder.

All of this would be impossible to achieve without the strategic intervention of artificial intelligence (AI) and machine learning (ML).

AI & ML Holds the Key to the Gates of Succes

One of the biggest challenges faced by P&C insurers is the limited nature of insights into a property at the time of new business generation or policy renewal. This is even more critical for commercial customers where the property value can go up to millions of dollars.

The insurance industry is also notorious for its poor quality of customer experiences – using chatbots, customer analytics, and automated interactions, P&C firms could augment their customer service capabilities for increased loyalty. Fraud detection and prevention is another benefit to be gained from AI & ML intervention. Fraud is a critical concern for insurers across the globe – for example, the UK witnessed one insurance scam every minute in 2018. This comprises primarily of fraudulent claims and dishonest applications, easily preventable by the use of AI & ML.

So, what are the different AI & ML technologies available today, and how could insurers gain from them? Here is a quick overview:

  • Data and textual analysis - This comprises the lion's share of use cases, with AI engines scanning personal information records and corporate audits to identify the possibility of risk. Based on the insights, P&C firms can customize their pricing policy or even deny insurance to an entity that it deems as a high-risk proposition. Today, with so much information available publicly online, insurers not using AI are clearly missing out.
  • Image/object recognition - This is particularly relevant for commercial P&L, during the time of policy renewal. Firms do not have to send a representative to be physically present on site and take photos and videos of the insured property, it can be evaluated using AI & ML to detect changes. This will avoid errors arising from human fatigue, which is a common occurrence for insurance executives assessing properties at scale.
  • Customer-facing chatbots - Customer support is a massive cost center for any business and this applies to insurance as well. Often, customers will insist on having a one-on-one conversation even if online documentation and FAQs are available. 24/7 customer support can positively impact customer loyalty levels – and AI & ML makes this possible without straining the workforce.
  • Robotic process automation (RPA) - Backend activities in insurance are prime candidates for AI & ML implementation, unlocking significant efficiency gains. Claims collection is a one-such process where AI-led RPA can bring in accurate and adaptive automation capabilities. Combined with ML, automation engines will become more effective over time, resulting in an incremental uptick in profitability.

Practical use cases: How is AI and Machine Learning is disrupting P&C insurance segment

Insurers can use AI & ML in some of these ways to change operating models for the better. Areas such as robo-advisors are still at a nascent stage, with plenty of promise in the long term.

Aligning AI & ML to Unique Industry Needs: The Way Forward

These technologies could play a role at six moments of truth for P&C insurance stakeholders. Here's how an AI-led insurance value chain would look:

  • Marketing/sales - AI can help hyper-customize marketing initiatives based on audience segmentation and sentiment analytics.
  • Insurance actuary - AI can create more accurate exposure analysis models, specifically in trends-based areas such as catastrophe modeling.
  • RPA in distribution - Processes like agent support, form filling, performance evaluations, etc., can be fully automated.
  • Customer servicing - As mentioned, chatbots can take over query resolution and help upsell/cross-sell to existing customers.
  • Claims management - FNOL reports can be analyzed by an AI engine to pursue investigation, curbing the risk of fraudulent claims.
  • Renewals/reinsurance - AI can assist in price optimization, negotiating the best value based on risk.
The 'new normal' in insurance value chain: six moments of truth for P&C insurance stakeholders

Clearly, AI & ML opens up a world of new possibilities for P&C insurance providers. America's leading firm Allstate is using this technology to shrink customer waiting times. Its new AI-enabled customer support mechanism can process an impressive 25,000 queries per month, reducing lost business volume. Examples like these demonstrate how AI & ML could be a revolutionary force – and the future is right here, right now. Firms that make bold, early moves stand to gain significant ground in the near future.

Share this page