AI perspectives from Tractable.

Back to blog list: See all posts

16 Nov 2017

Together Towards AI: Notes from InsureTech Connect 2017

The biggest names in insurance descended upon Las Vegas for the InsureTech Connect conference this October. The Tractable team was on-site, with CEO Alex Dalyac speaking on a panel about the future of artificial intelligence (AI) and insurance. Alex discussed how cutting edge computer vision can now automatically assess auto collision damage and determine costs. Tractable’s work with firms like Ageas, Mitchell International and some of the largest US auto insurers, will bring the promise of AI to millions of policyholders.

With over 25 million estimates and $60 billion of repairs in the US annually (45% of which are not visually reviewed), the impact of Tractable’s AI on the insurance industry will be significant, both in operational improvements and in a better customer experience. New technology jobs will be created and insurance estimators augmented by AI can spend more time focusing on the customer. Universal safe and efficient repair standards based on AI could finally bridge the gap between insurers and repair shops, reducing miscommunication and producing more accurate estimates. Consumers will feel the impact – the day when you can take photos after an accident and receive insurance quotes immediately on your smartphone is not far away.

The Tractable team spoke with a number of insurers and technology companies that are working to make AI augmented insurance a reality across multiple applications and use cases. Speaking to thought leaders, business leaders and technologists, it is clear that these various stakeholders will need to work together to implement AI into the insurance process.

“AI allows for the passive gathering of information on driving so that when you have that next moment of interaction with your insurer, they can tell you something useful because they have been learning all along.”

Farooq Sheikh, Principal at Oliver Wyman

“AI in insurance will allow carriers to deliver scalable and customized solutions for members and policyholders,” says Ramon Lopez, Vice President of Property & Casualty Claims and Innovation at USAA. “The challenge comes from translating the questions and problems within insurance into actual solutions. There needs to be a way to alleviate friction, but there also has to be interaction between all companies and internal teams.”

This begins with defining the problems that AI can solve, which are multifaceted. Farooq Sheikh, a Principal at Oliver Wyman focused on financial services, describes it as a “toothbrush or dentist” sort of problem. The question is whether insurance should be a day-to-day concern for consumers – like the regular maintenance of brushing your teeth – or whether it should be an annual concern, like visiting the dentist for a regular check-up. “We talk about moments of interaction and moments of influence,” says Farooq. “These can be acute, say when you speak with your insurer after an accident, or they can be constant, such as interacting with an app on your phone. AI allows for the passive gathering of information on driving so that when you have that next moment of interaction with your insurer, they can tell you something useful because they have been learning all along.”

In building AI-driven solutions to various insurance issues, insurers are taking very different approaches. “Insurers are either in denial, looking to buy AI startups, or seeking partnerships,” says Farooq. “Those partnerships may echo the model that we see in pharmaceuticals, with insurers completely outsourcing their R&D, or we may see insurers learning from their startup partners and developing their own capabilities alongside them.”

Ramon Lopez of USAA says: “Insurance companies are increasingly partnering with startups, leveraging collaboration as opposed to a transfer of ownership. Working with InsureTech startups allows for easy integration of products within insurers. The relationship and coordination is quite simple. But lots of work must be done upfront to figure out the exact problems to solve together. The main thing is to be in a position where all the right puzzle pieces fit together.”

“Working with InsureTech startups allows for easy integration of products within insurers.”

Ramon Lopez, VP of Property & Casualty Claims and Innovation at USAA

As Will Meneray, a principal in the Oliver Wyman digital practice, puts it, “After a firm agrees to work with a startup – once the immediate rush of excitement subsides – there’s often an awkward silence when both sides realize they don’t really speak the same language. You will never realize the potential of AI that way; you need to find a way to integrate the two cultures and ultimately work together.”

For many established players, the push to innovate is internal; they are well aware of issues with their current business and operational models. Insurers will certainly worry about how new services made possible by AI may cannibalize revenue from their traditional lines of business. Paul Rosenstein from Mitchell International, a technology provider to major insurers, says: “Insurers will be methodical. They know that they need to evolve, but also that it doesn’t make sense to rip out large parts of their business all at once.”

This has implications for where AI will fit within the insurance ecosystem. Kayvon Deldar from Plug and Play Tech Center, an accelerator in Silicon Valley, believes that AI will be used to solve the necessary day-to-day tasks of running a business, such as the automation of routine patterns or to augment CRM, rather than for large, flashy use cases.

Others believe that consumers will lobby for innovation given the benefits they will see elsewhere in service delivery. It may come down to how well AI is able to tailor solutions for individual customers. “AI has already proven successful at facilitating personalized responses to situations, such as providing insights into the type of accident that an individual is likely to be involved in,” says Ramon Lopez of USAA. “Consumers will come to expect this personalization, and insurers will continue to use AI to deliver it.”

Back to blog list: See all posts

Next post: How to Evaluate AI in 5 Minutes

Get updates