Building a successful AI business case in insurance

The pandemic has not been kind to the insurance industry. This unexpected global disruption caught insurers off guard, with 87% of those responsible for insurance operations saying it exposed shortcomings in their organization’s digital capabilities, according to Deloitte’s 2021 Insurance Outlook. Additionally, the pandemic’s financial impact has been significant. Expectations are that, for 2020, global non-life premiums will have been flat, with a 1% decline in advanced markets.

But the long-term trends have been challenging, even without the pandemic. Investment yields have sharply declined while combined ratios have increased steadily over the past three years. Insurance leaders are eager to find ways to become more efficient, and, in response, 95% of insurance executives say they will accelerate their digital transformation efforts.

Expanding the use of AI

Artificial intelligence (AI) plays a significant role in these digital transformation projects. In Europe and Asia-Pacific, expanding the use of AI in underwriting is their number 2 priority, though North America placed a much lower emphasis, ranking it as number 8. That’s not necessarily surprising because while AI has the potential to powerfully transform and optimize insurance processes, deploying and leveraging it effectively requires a careful approach with a great deal of planning and forethought. After all, 80% of companies report that their AI projects are stalling, according to Dimensional Research.

That said, so long as IT works closely with operations and the business in a deliberate, considered manner, AI can have a powerful impact on a carrier’s business. To ensure that a carrier’s initiatives meet and exceed expectations, insurance leaders should make sure they incorporate the following into their AI planning process.

Identify business objectives

Management must have a clear idea of what it hopes to accomplish at the very beginning of the process. Deploying AI simply because everyone else is doing it is a recipe for failure. Typically, AI is used to gain efficiencies by automating and optimizing one or more of the following:

• Reducing leakage: Loss is the top expense item for insurers, and premium leakage is estimated to amount to billions of dollars annually. The number-one expense item for insurers is loss. Verisk estimates that personal auto insurance leakage alone amounts to a $29 billion problem. AI can help plug these leaks by detecting fraud patterns, over or unintended exposures and misalignments.
• Augmenting underwriting: Risk evaluation is typically a manual process using third-party reports. AI can rapidly extract risk factors to produce an evidence-based grade, which enables risk engineers to focus their time on complex cases. As a result, underwriters can produce quotes faster, providing a competitive advantage.
• Accelerating claims: AI can automate the review of claims packages to determine their level of complexity and route them appropriately. They can extract data from claims to assist with liability determination and provide suggestions based on that data, automatically review claim packages to evaluate complexity and route them accordingly. The result is an optimized, streamlined workflow that enables claims handlers to make faster, more informed decisions, which can significantly improve customer satisfaction.

While working towards multiple objectives is possible, be careful. It’s easy to become too ambitious, especially given the powerful impact that AI can have. The safer strategy is to focus on a single objective and use case, learn from that experience, and then, once it has proven a success, expand the use of AI within the organization.

Finding the right use case and KPIs

With the business objective identified, the next step is to determine where in the organization AI can provide these benefits. For example, if the goal is to reduce the cost of service, analyze your lines of business to establish which one is both experiencing the most pain and providing the greatest opportunity for AI to deliver value.

As a guide, consider the following two metrics in your valuation: 1) How many transactions or cases, such as claims or policies, does this particular function handle, and 2) What gains in terms of efficiency, such as hours saved per claim, are likely?

Follow the numbers. Political pressure and other factors within an organization may push IT to implement AI in less than optimal use cases. Again, bowing to this pressure when the data points in another direction will likely lead to disappointing results. Especially with relatively new technology such as AI, it’s critical to achieve success early to gain confidence and, just as important, experience so that broader applications will not only have widespread support; they’ll be more likely to achieve expected results.

And, speaking of results, management will need to establish tangible KPIs to measure and demonstrate success, which can be tricky to do for AI. That said, there are some best practices for carriers to follow.

Look at reduced exposure. Leakage is a pressing issue for most carriers, and through the use of AI in risk engineering, risk assessments can take place faster, and the grading system can become more consistent. Efficiency is another good metric to evaluate, especially for underwriting processes. AI can significantly cut the amount of time that it takes to process documents such as policies or risk evaluations, which can boost output and increase capacity. Process efficiency doesn’t just cut costs; it can improve and accelerate customer acquisition and growth, so that’s another good KPI to track. Faster policy turnaround puts an organization in a much better position to close business.

Once the above items are accomplished, it’s time to map out the infrastructure. AI isn’t a plug-and-play technology. It needs to be integrated tightly into existing processes, so the AI implementation team will need to carefully determine what expertise, data and other systems will be required.

Insurance faces a challenging future, and AI can help companies meet those challenges, but only if they approach AI strategically and thoughtfully. With a focused deployment that produces early success, carriers can position themselves for greater gains in the future with this exciting new technology.

Pamela Negosanti is the head of sales and sector strategy for financial services and insurance at expert.ai. Contact her at pnegosanti@expert.ai.

Related:

• Using technology to disrupt the auto insurance supply chain
• Are your insurance processes tech-enabled or tech-shackled?
• Lemonade’s tweet causes uproar over its use of AI