Transforming the insurance industry with AI

The insurance industry has not been slow to adopt digital technology. More recently, though, artificial intelligence (AI) has transformed the industry further by enabling systems, that are already fast and mobile-ready, to be enhanced with smart technology. AI technology uses Big Data – the processing of very large volumes of many types of rapidly changing data – and machine learning (ML) to enable systems to improve continuously, without the need for active human intervention.

 

AI is a critical differentiator in the insurance industry. By analysing vast amounts of data, including risk factors and customer information, AI algorithms can assess risks in real time, determine coverage eligibility and set appropriate premiums. By streamlining the underwriting process, efficiency is improved and commercial success is augmented.

 

But AI doesn’t just improve internal processes. It is at the heart of changing the relationship insurance companies have with their customers. Insurers are increasingly adopting AI technologies to enhance the customer’s experience. This includes offering personalised insurance products, enabling customers to manage their own policies and claims (“self-service”) and delivering more effective communications across all channels.

 

Personalised products

 

Rather than basing products and pricing on groups of customers, for example by postcode or other demographics, insurance companies are now far better able to provide products and pricing targeted at individual customers. By automatically analysing detailed personal data, individual preferences and behavioural patterns, they can provide customised products, personalised pricing and targeted policy recommendations, and do so in seconds rather than having to wait for human intervention.

 

Data analytics and AI can be used to assess individual risk profiles. Personalised products can then be offered, where competitive pricing is based on an evaluation by AI of the likelihood of claims being made in the future. Historical claims data can be combined with current and recent personal data, such as lifestyle or driving habits, to provide more accurate risk assessments, enabling competitive quotes to be offered to low-risk individuals.

 

AI also allows the easier customisation of policies without the need for expensive assessment by human underwriters. As a result, flexible coverage options become easier to provide and it becomes possible to price new insurance products, such as cyber-security or identity theft protection, on an individual basis.

 

Another advance facilitated by AI is the provision of usage-based and telematics solutions. By using telematics devices – for example, in cars or even just mobile apps – to collect data on customer behaviour, companies get the ability to monitor and analyse customer activity and can make automated pricing decisions accordingly.

 

For example, a car insurance company might offer variable monthly insurance premiums based on driving habits such as speed, mileage and braking patterns. A trip to the seaside for a week might increase insurance costs for that month based on the extra miles driven, but the cost per mile for the insurance might be less because the customer’s driving was safer than in a normal month where they commute to work through heavy traffic.

 

Based on real-time performance data, insurers can adjust premiums, encouraging safer driving practices and rewarding responsible behaviour.

 

Companies could even offer personalised feedback to policyholders. This could include recommendations on how to mitigate specific risks or suggestions for improving safety and security. By sharing insights and guidance, insurers could help customers protect themselves and their assets better. This type of proactive risk management advice is something that would benefit both the customer and the insurance company.

 

Enhanced customer experience

 

The insurance industry does not always succeed in being appreciated by its customers. Claims handling in particular is a critical touchpoint that can leave customers frustrated and angry at a time that they are under unusual stress.

 

Insurers can use AI to expedite the claims process, reducing this stress. Currently, “self-service” is commonly used as a way of saving costs and of enabling the customer to have some degree of control over their claims. But self-service can also add to frustration when the interactive voice response (IVR) systems that are typically used fail to cover all eventualities, or when a customer simply wants to talk to another human being.

 

AI systems are far better able than fixed voice recognition systems to respond to complex customer queries; they can get even better over time as they learn about unusual circumstances. This is not just because IVR systems are not always very effective at “understanding” the sentences that customers use. It is also because the number of options IVR systems can deliver is limited by practical consideration.

 

AI-powered chatbots potentially have no such limitations and can respond to a wide range of queries (a range that will grow wider over time) because they can interpret the intent behind questions rather than simply responding to keywords. Because they use natural language processing to understand customer queries, they can deliver relevant responses. And they are less likely to be thrown by accents and slang that may not have been programmed into an IVR system.

 

Another important benefit of AI-powered chatbots is that they are able quickly to pick up on emotional clues that indicate a particular customer would be better served by a human rather than a machine.

 

Document upload and management can also be enhanced by AI. Insurance companies often provide secure document upload functionality through their online portals, where customers can submit required documents such as proof of loss, photographs of damages or policy-related documents directly through the self-service platforms. AI can add a layer of quality control here, identifying whether the information submitted by the customer is likely to be sufficient, and indeed looking for instances of potential fraud that can be highlighted.

 

Effective communication

 

Insurance companies are beginning to use AI to improve the way they communicate with their customers, shifting towards proactive communication strategies powered by predictive analytics.

 

Changes in behaviour and lifestyle can be used to provide relevant insurance product offers or targeted messages about policies. For instance, the change from driving a sports car to driving an SUV might signal a change in circumstances – perhaps a new addition to the family – that could represent an opportunity to sell different products such as life insurance or health insurance.

 

Another important change to communication that AI facilitates is the ability to manage customers and prospects across different communications. People expect a seamless experience across the various channels they use, including online, mobile and in-person interactions. Insurers that adopt an omnichannel approach can provide a consistent level of service that recognises who the customer is, and what information they have already shared, whatever channel is being used.

 

Seamless transitions across channels, where the customer’s context and history are carried forward during channel switches, and where appropriate information can be offered based on an analysis of their history and current circumstances, will play a major part in reducing the frustration often experienced with more traditional systems.

In addition, AI can be used to analyse omnichannel journeys, mapping out customer journeys across different channels to understand how customers move between touchpoints and the challenges they may encounter as they do so. Pain points can be identified so that ML can help to optimise processes and create a seamless experience over time.

 

Managing the future of AI in insurance

 

There are many ethical issues around the use of AI in insurance. Accountability is one – a human must be accountable for decisions made by an AI system and stakeholders must have a clear route to challenge decisions and, where appropriate, gain compensation.

 

In addition, the issue of fairness is frequently raised as a problem with AI, an issue generally caused by the use of inadequate and incomplete data to train the algorithm’s decision making. Another problematic area is the use of personal data in ways that data subjects may find unexpected or unacceptable. And some potential features of AI, such as the use of emotion-tracking to influence decisions, may fall foul of new legislation..

 

But despite these issues, AI holds huge potential for the insurance industry. Understanding what that potential is and how it can be embraced legally and ethically will be key to the future success of the industry.