Artificial intelligence use cases explode in insurance
Artificial intelligence is a term for a very broad array of technologies that mimic human cognition and activities; they can also discover patterns and relationships that go beyond anything humans are capable of. Depending on your view, AI will either be a great boon to human society and business, or an existential threat to humanity. Or you may believe that the whole area is overhyped and won’t have these dramatic implications. Whatever you believe, it is important to understand how AI applies to the property/casualty industry and where the greatest potential lies for harnessing the technology.
A new research brief by SMA, based on a survey of insurance executives, provides some insights into these areas. AI in P&C Insurance: Potential and Progress covers personal and commercial lines, revealing significant differences between the sectors. AI has potential in P&C to address many business issues across the enterprise for every sector of P&C. Today it appears in use cases here and there. Over time, AI will contribute to solutions everywhere in P&C.
Insurers are experimenting with and implementing AI technologies such as robotic process automation (RPA), chatbots, data and text mining, and machine learning. Underwriting rules engines and solutions for claims fraud are being enhanced with newer AI capabilities. Underwriting and claims are two areas that have been using earlier forms of AI (case-based reasoning, rules engines) for some time. These areas still offer great promise for using AI in the future, but now AI is being applied to customer-facing areas as well as other operational areas (e.g., marketing, distribution, policy servicing). For example, insurers are now using AI technologies to improve the customer experience – in fact, personal lines insurers see that as the area to reap the most value from AI overall. Commercial lines insurers tend to expect more value from a better understanding of risk and more efficient operations.
It is important for insurers to actively investigate AI technologies and how they might apply to strategic or operational business needs. What makes AI so important and applicable to many insurance use cases is the range of technologies that are part of the AI family. Depending on how you group and count them, there are at least a dozen different AI technologies. In addition to those already mentioned, there is image recognition and visioning systems, natural language processing (NLP), cognitive computing, artificial neural networks, and others.
Over time, various AI technologies will become embedded in many different solutions and be used across the enterprise. Now is the time to explore, experiment, and look for alignment to business strategy where advantage can be gained.
Note: For a comprehensive view of the AI landscape for insurance, including definitions of many AI terms, see the SMA Research report, AI/Machine Learning in Insurance: A Force to be Reckoned With, published in April 2017.
This blog entry has been reprinted with permission from SMA.