The Future Of AI In Insurance

Artificial intelligence (AI) was the phrase on everyone’s lips and featured on almost all the top trends lists as we entered the new decade. AI is a computer-based science that mimics the perception, reasoning, learning and problem-solving of the human mind, drawing on the vast swathes of data that we are all now producing courtesy of our ‘connected’ lives.

Home assistants, smartphones, smart watches, connected clothing, eyewear, home appliances and medical devices… all these gadgets gather a huge amount of data that can be used to improve and enhance our lives, all courtesy of artificial intelligence and machine learning.

How is artificial intelligence used in insurance?

Artificial intelligence is expected to alter many facets of the insurance industry, and is already making an impact in many areas. From ultra-tailored policies and pricing based on precise, data-powered risk assessments to automated screening of fraudulent claims, the future seems set to enjoy boosted productivity, more streamlined decision making, lowered costs and an enhanced consumer experience.

Let’s look at some of the specific areas of the industry where artificial intelligence is transforming insurance.

Behaviour based insurance

Artificial intelligence will take the concept of behaviour based insurance several steps forward. Instead of tailoring a policy to the information provided by a customer, access to connected devices such as vehicles, fitness trackers and voice assistants will assist insurers in understanding their customers’ needs and behaviours at much deeper levels. This will lead to highly personalised policies, as well as a move for insurers to create a greater range of product offerings.

Personalised policies mean personalised pricing, which in theory should be fairer for the policyholder; for example reduced premiums for safer drivers.

AI allows policies to be created in significantly shorter times too, meaning customers in the future need not wait more than a few minutes for cover to be put in place. In fact, the age of ‘instant insurance’ is likely soon to extend to more complex areas of cover that would usually take time to assess.

The benefit of using AI in analysing the data presented by connected devices is that the processes learn and adapt over time. This ‘machine learning’ keeps pace with evolving customer situations and behaviours, as well as external changes like environmental factors. Eventually, such changes will become readily predictable, providing a golden opportunity for insurers to proactively develop and offer highly relevant products that suit specific needs.

Underwriting

Accurate underwriting comes only from appropriate risk assessments. Assessments are traditionally based on information provided by the insured and then manually cross-referenced against available information.

This is a lengthy process that has high potential for error. Artificial intelligence will help transform this process by bringing together information from a wide scope of sources including social media and public databases, providing a faster and more thorough view for a greater precision risk assessment.

Fraud detection

More than half a million insurance fraud cases were reported in 2017 according to the Association of British Insurers (ABI). A number of top insurers are already employing AI technologies to manage fraud detection.

There are two main types of AI-based fraud detection. One is anomaly detection which uses machine learning to first establish what a regular claim would look like, then applies this as a benchmark to compare against. When a claim deviates from the norm, the software flags it to a human agent to decide whether to accept or reject the notification. This ‘accept or reject’ action itself further boosts the learning of the technology, helping it understand whether it is drawing the right conclusions.

The other type of AI fraud detection technology is predictive analytics. This involves claims experts manually labelling a large number of fraudulent and legitimate claims. The machine learning then steps in to recognise fraud methods used in the labelled fraudulent claims.

Claims management

Artificial intelligence has the ability to considerably speed up the claims management process. Insurers are traditionally faced with processing reams of information when a claim is made, most of it manually.

It is a lengthy process which can result in policyholders waiting for some time to see their claims settled. AI and machine learning are able to automate much of the claims procedure, right through from the initial reporting process to validating the claim and communicating with the insured party.

Looking further into the future

Artificial intelligence is already shaping the insurance industry and, as consumers, brokers and insurers become more at one with advanced technologies and their use further increases, so change will accelerate. The ultimate outcome will be increased productivity and reduced risk for insurers, and a much enhanced experience for the consumer.

Insurance will shift from ‘detect and repair’ more towards ‘predict and prevent’. AI could in the future allow insurers to provide their policyholders with advisory services such as safer driving routes, or by flagging early signs of damage in the home that could lead to a claim situation. AXA for example already offers policyholders its ‘Xtra’ health app which includes a chatbot that makes personal suggestions as to how nutrition and fitness goals can be met.

Possible downsides

Critics argue that AI could result in negative outcomes for consumers, especially where large volumes of data are being collected for analysis. This could impinge on privacy if consent is not given. There is also concern that ultra-personalised risk assessments may render some individuals ‘uninsurable’ by revealing risk indicators that were previously undetected. There is also potential for discriminatory factors to come into play, such as in situations where individuals with chronic health conditions find themselves unable to afford premiums.

However, the flipside of using artificial intelligence in undertaking risk assessments is that some insurance products may become more accessible to individuals who would previously have been considered too high a risk.

Providing the industry commits itself to universally acceptable levels of privacy and data usage control so that AI is used in a positive rather than negative way, the future should be a better one across all areas of insurance thanks to artificial intelligence.