Preparing your business for artificial intelligence (AI)

Whether it’s “Blade Runner” or “2001: A Space Odyssey,” you can count on audiences to devour dystopian works about artificial intelligence (AI).

In spite of science fiction’s dire prognostications, real people believe AI will make the world a better place. It follows that insurance industry leaders are jumping on the AI bandwagon. Out of the $1.7 billion invested in InsurTech last year, more than 40% was aimed at AI.

What’s more, Accenture’s Technology Vision for Insurance 2017 survey of more than 550 insurance executives from 31 countries found that 75% of respondents believe AI will significantly or completely transform the insurance industry over the next three years. Two-thirds of those polled by Accenture said they expect AI to significantly or completely transform their own companies.

Brave new word

AI is loosely defined as the ability of a computer to mimic various types of human behavior — whether that’s learning, speech and language, vision and perception, or analysis. It includes emerging technologies like cognitive computing, machine learning and natural-language interfaces.

Although rudimentary versions of AI emerged as early as the 1940s, AI technologies are only now bursting onto the business scene. For insurance, that means we now have vast troves of data, including Internet of Things (IoT) sensor data, which can drive advanced machine learning and deep learning algorithms to look for patterns. Consider that by 2020, the average new vehicle will contain as many as 200 sensors, including sensors for the engine, body, chassis, global positioning system (GPS) functions, seatbelt and tire pressure.

Converting to digital models

Faced with current top-line growth and threats from competitors adopting emerging digital technologies, insurers are looking to AI to transform their business. They’re attempting to move from being process-driven to customer-centric, thanks in large part to the preferences of digitally-native millennials, who are now entering their prime insurance buying years.

AI is allowing insurance companies to improve the customer experience in such basic processes as product design, product sales and consulting, risk assessment, claims processing, basic fraud detection, pricing and underwriting.

Consider that one AI InsurTech startup plans to offer digital solutions for buying home, auto and business insurance online. Another recently raised seed money for AI to help consumers re-shop their insurance automatically.

AI advocates predict that chatbots will evolve into “robo-advisors.” RiskGenius, which analyzes language in claims policy documents, and PolicyGenius, which compares insurance quotes, are examples of AI moving in this direction. Robo-advisors in the broker space, predicts PwC, will soon offer “recommender” systems and “someone like you” statistical matching. Mature robo-advisors will someday understand balance sheets and income statements, and be able to recommend, monitor and change financial goals and portfolios for policyholders.

Risk assessments stand to become far more accurate. In the automobile insurance market, for instance, using sex, age, experience, driving record, number of annual miles and locations was once the best actuarial model available. But monitoring actual driving behavior is even more predictive of risk. With vehicle sensors producing telematics that can be collected in real time, that model is now possible.

Claims processing also will become more automated. Cars will be able to self-report minor accidents. By using AI to process images of an accident, the system automatically can determine its severity and decide whether it should be written off or whether the car should go to a service station.

At the same time, claims organizations can use the same AI technologies banks use for identifying fraud in credit card transactions to identify claims fraud.

Preparing your business for the journey

How can the insurance industry become fully prepared for AI? At a strategic level, companies must be prepared on three fronts: Process, people and technology.

Process: The biggest challenge most insurance organizations face in moving from small AI projects to embedding AI innovation throughout the organization is enforcing the necessary cultural change. Insurance firms must find ways to deploy innovation into very traditional processes — transitioning from being a process-driven to pattern-driven organization. That’s a big change and it will take time.

People: Companies must also manage fears regarding job security. Rather than wipe out jobs, AI will enable people to work on sophisticated tasks while machines automate more trivial processes.

Because data science is a nascent field academically, producing fewer than a couple hundred graduates a year, organizations will need to use this scarce talent efficiently and effectively.

Technology: Not only will new hardware and software be required, insurance companies will need a technology platform that can nimbly accommodate changing business strategies and priorities while ensuring control over data use, security, redundancy, replacement planning and provisioning.

At the same time, companies must overcome the danger of creating analytics silos — islands of analytics technology, data and skillsets — by implementing a collaborative environment that allows people to work together.

To make the most of AI, insurance companies must take a hard look at their people, processes and technology and find ways to innovate and implement these solutions at scale. The result promises to be a digital transformation that will bring customers more of what they want and need with the experience they demand.

David Hartley (David.Hartley@sas.com) is global director of Fraud & Security Intelligence at SAS. Norman Black (Norman.Black@sas.com) is EMEA Insurance Industry principal at SAS.

See also:

4 emerging risks in the world of artificial intelligence

5 artificial intelligence tools defining the future of P&C insurance