Where Are They Investing and Why?
In a recent study of 13 different verticals, the insurance industry invested more on artificial intelligence (AI) than any other industry—on average, $124 million per company surveyed.1 Of the 54 insurance companies that participated, about half were in North America, and about half were in the property and casualty industry.2 While the greatest areas of investment were security and customer service, investments were made across multiple business functions, including everything from human resources to sales.
Where insurers are investing is interesting, but I’m also fascinated by why the insurance industry is investing and the types of business challenges they are looking to solve. Near term, the industry is investing to improve operations and to make better, more informed decisions around claims, but long term, artificial intelligence has the potential to transform every aspect of the property and casualty and collision repair industries. Eventually, it may not only help us make better decisions, but also deliver insights that we’ve never seen before.
AI Today—Making Better Decisions around Claims
Many of today’s artificial intelligence applications are focused on achieving operational efficiencies, both in customer-facing interactions and behind the scenes. It is used for everything from automating repetitive tasks to identifying fraud. One of the most widely adopted applications for AI in insurance is chatbots. According to Gartner, chatbots will power 85 percent of customer interactions by 2020, and the average person will have more conversations with a chatbot than with their spouse.
A big challenge with this type of AI is its low EI, or emotional intelligence. While chatbots can ask and answer questions, they are not great at reading emotions or understanding tone. A growing field of AI study called sentiment analysis is changing that. Sentiment analysis, sometimes called emotion AI, analyzes written or spoken words to understand the feelings behind them. Intelligent solutions like Watson Tone Analyzer are using it to help chatbots understand emotions and interpret tone and are a big step toward making these interactions more human and personal so that they better serve the customer.
As chatbots grow more sophisticated, they are moving beyond customer service and into other operational functions where they can better serve insurers as well as their customers
As chatbots grow more sophisticated, they are moving beyond customer service and into other operational functions where they can better serve insurers as well as their customers. Natural language search—similar to search on Google or Bing—is beginning to provide the enterprise with straightforward access to their data without complicated query methods. In doing so, information that was once only available to a trained user who could pull a report will soon be available to anyone with a question. What does this mean to insurers? As natural language search gains traction, information is becoming more accessible and that information can be used to inform decision making around claims—and both insurers and their customers benefit from that.
AI Tomorrow—Delivering Unprecedented Insights
Beyond natural language, there is a growing multitude of ways AI can deliver information and recommendations so people can make well-informed business decisions. We’re doing this at Mitchell with WorkCenterTM Assisted Review, a solution that uses AI to validate repair vs. replace decisions for damaged vehicles. To train the AI, we uploaded millions of photos of damaged vehicles across all makes and models of cars and trucks. Alphabet’s DeepMind used a similar process to train their AI entity AlphaGo to play the ancient game of Go using thousands of professional and amateur games.
Although the rules are simple, Go is infinitely more complex than chess—the number of board configurations is 10 to the power of 170. Mastering it was considered one of the foremost machine learning challenges. Unhindered by preconceived human notions of the best ways to play the game, AlphaGo upended hundreds of years of conventional wisdom by making a number of innovative moves to beat one of the best Go players of the last decade, Lee Sodol, at his own game.
A later version of the AI, AlphaGo Zero, bypassed the training step and learned to play entirely on its own using a technique called reinforcement learning. In just three short days, it taught itself how to beat the original program.
AlphaGo provides insight into where AI is headed. You can see the opportunity to dramatically reshape how work gets done. Like AlphaGo, AI may even be able make innovative “moves” that depart from conventional thinking and result in faster, more accurate, and more economical claims and collisions resolutions.
AI in Property and Casualty—Restoring People’s Lives
In the Tata study, participants were asked to rank what they thought the biggest risk was to successful AI implementations—an interesting question when posed to an industry built on evaluating risk. The number one answer: developing a system that makes good, reliable, safe decisions.3 To me, this is also the area of greatest potential reward.
Our greater purpose as an industry is to restore people’s lives after an unforeseen, and often challenging, event. In our role at Mitchell, that means providing solutions and services that support the proper and safe repair of vehicles, and help people get back to their pre-injury state after they’ve been injured in a vehicle or workplace accident.
We’re already beginning to reap the operational benefits of AI, but for me, one of the most exciting things about artificial intelligence, and many other forward-looking technologies, is how the power of data will bolster human decision making to simplify the inherent complexities and uncertainties of restoring people’s lives.
1 TCS Global Trend Study Phase 2, Getting Smarter by the Sector: How 13 Global Industries Use Artificial Intelligence, Tata Consultancy Services, Page 15, Accessed 2/18/2018
2 TCS Global Trend Study Phase 2, Getting Smarter by the Sector: How 13 Global Industries Use Artificial Intelligence, Tata Consultancy Services, Page 54, Accessed 2/18/2018
3 TCS Global Trend Study Phase 2, Getting Smarter by the Sector: How 13 Global Industries Use Artificial Intelligence, Tata Consultancy Services, Page 57,Accessed 2/18/2018