The potential impact of Artificial Intelligence (AI) upon various professional sectors has sparked off lively discussions over recent years. But despite doomsday predictions of computers taking over the world, at present AI is predominantly being developed for businesses to improve efficiency, increase automation and - perhaps most importantly - more effectively manage and analyse big data.
How can AI be used by insurers?
Policy customisation and sales
Insurers hold vast amounts of data on their customers but much of the inherent value of this data remains untapped. AI software can help to organise, cleanse and refine this mass of data, applying analysis to discover trends which can be used for various purposes, including:
- Predicting customer behaviour and assess risk, enabling the customisation of individual policies and premiums
- Assessing customer lifetime value
- Discovering latent upselling opportunities
- Suggesting new types of insurance products according to trending demand
- Telematics – in the case of car insurance, a “black box” can be fitted to the vehicles of policyholders, monitoring driving style and patterns and providing analytics data to the insurer, providing lower renewal rates for safer drivers
Chatbots have become commonplace on many consumer facing websites. They are often used primarily as a preliminary customer service function, acting as a customised search function, directing the user to the most relevant webpage or providing them with the most relevant contact details. Their ability to process natural language is crucial to their effectiveness.
Aside from using chatbots - also known as robo-advisers - as receptionists, insurers can suffuse them with more advanced AI which allows them to answer basic questions about a policyholder's account (eg next renewal date), inform them about relevant products and even help with policy applications and claims processing. One digital disruptor, Lemonade, offers two bots - Maya and Jim; the former creates personalised insurance policies whilst the latter processes claims.
Fraud prevention and claims processing
The ability of AI algorithms to sift through big data and spot trends, patterns and anomalies, can also help insurers to detect fraud, both in terms of suspicious claims and also questionable policy applications. Manual tasks of combing through reams of information pertaining to a claim can be automated, leading to significant savings in time and cost.
Fraud detection aside, claims processing which uses AI can lead to greater efficiencies. Digital disruptors who have implemented touchless claims – where “technology is used to report a claim, capture damage or invoices, run a system audit and communicate with the customer electronically" - have already witnessed a reduction in cycle times. Adding blockchain technology and smart contracts into the mix can even allow payments to be automated for certain types of claim, as can be seen with AXA's current trial of new flight delay insurance product Fizzy.
How can insurers prepare for the future?
Although savings can be made by effective implementation of AI, the increased reliance upon software algorithms by insurers also ramps up the potential impact of a cyberattack. Introducing touchless claims can streamline processes when technology is working as intended - but AI which "goes rogue" as a result of a hack can wreak havoc, so insurers will need to pay more attention to their cybersecurity.
As well as paying heed to their own cybersecurity, insurers will be increasingly asked to provide cyber insurance policies to address the reliance upon AI by the wider business community. Particularly sophisticated types of AI may require policies all of their own, and new questions will need to be tackled in the claims process: who is ultimately responsible for the decisions taken by a self-thinking software algorithm and where does fault lie for mistakes (eg. if a driverless car has to decide whether to crash into another car or veer into the pavement when a crash becomes inevitable)? What is clear is that insurers will need to keep up to speed with the latest advances in AI and adapt their policies accordingly.