As an AI developer for the insurance industry, we have been asked how we see the future of AI a few times recently, so we thought it may be worth writing a blog.
To start with, the future of AI generally will probably be different to the future of AI in insurance for many reasons. We believe the main reason is that insurance is a regulated market, unlike retail e-commerce for example. Being a regulated market may have an impact on the insurance industry’s ability for fast adoption, especially if the regulators want to get involved with the use of AI within the industry. So far, the regulation of AI within insurance is not present, however, with the enhanced use of generative AI this may change – especially if some of the current flaws of large language models (LLMs) are not solved, one of these being hallucinations (generating false information).
Anyway, let’s consider a user example as to how we think our tech can evolve and deliver service in the longer term. We assume we are talking about the year 2026 here and assume development and investment in AI continues at the same rate as it is today. I think it will be difficult to predict where AI and insurance will be beyond 3 years based on the way things are currently going.
Our use case story board
You are buying an insurance policy online though your phone.
An AI agent is talking with you about your risks, your concerns, what you’re worried about and what you would like mitigating. The AI is doing this while analysing your individual data for a personalised experience. The AI understands your local country economy, the social risks you face, and your job sector. The AI agent carefully crafts the policy, which is adapted to your personal situation, your risk profile and, importantly, the policy provides you with the benefits that suit you best. As you and the AI create this policy, you can understand it and you know there are no misleading clauses, loopholes, or funny words that only an underwriter know the meaning of. Its real time customisation of your basic policy.
The AI authorises the policy and instantly informs you the policy is bound and incepted – you are covered! Through your desired form of communication, the AI sends you the policy in a downloadable format, key policy benefits, along with who to contact and how if there is an emergency the AI is not already aware of. The AI also explains what it is monitoring for the enforced policy, such as weather, floods, your financial and personal data on the dark web, and so on.
The AI also proposes an optional extra monitoring, which is inclusive in the policy, where the AI will continue to monitor social and environmental factors that may impact you. This is your risk assistant – MiAssistant
You agree for the AI to continue to monitor your data. This is beneficial for both you and the insurance company as potential risks are known and informed as they arise.
One of your concerns is weather, specifically, hurricanes as you live in Miami. The AI detects early signs of a potential hurricane heading your way and potentially, it could be a big one. The AI warns you and credits your account with $1000 so you can evacuate from the area – a key part of the policy. The AI suggests 3 potential routes and 3 different locations using 3 different forms of transport. You choose to take a flight to Michigan where there is no current risk, and you have friends. Using the $1000 you book the flight and a hotel in Michigan the AI suggested and keep out of the way of the potential hurricane.
More news comes in via the AI – it’s bad news. The hurricane is hellish, it’s a category 4 and is going straight over your neighbourhood. The AI is tracking the hurricane and sending you updates. The AI knows your home address, and unfortunately the hurricane has gone straight over and through your house. The AI is simultaneously reviewing the news and tracking social media for updates and gain information on your local area. The AI identifies your area and has confirmed there is serious damage to your home, the garage is flooded, and its contents ruined including your car.
While doing this, the AI is also speaking to the insurance loss adjuster who agrees with the AI decision to accept that there will be damage to the property based on the evidence it has presented. The decision has been made to credit you with part of the agreed total benefit amount prior to physically surveying the property. You’re credited $25,000 plus additional travel and hotel costs. You arrive back home, and your house is a mess, total chaos. You call the AI, and you turn on your phone camera and GPS location. The AI knows you’re at your property and can see your house and the garage, out buildings and part of the 2nd level has serious damage. It can estimate the repair costs, there and then, the AI credits the remaining $50,000.
In this hypothetical scenario, there is a happy customer, no arguments, no debate with the insurance company, and no prolonged waiting times for money to help the owner out. The insurance company has not needed to send a surveyor to the property and has only had to involve a human once. This has saved the insurer money through AI and not having to involve humans. By providing this outstanding service the insurance company has gained a lifelong customer and a customer who will be talking about their fantastic experience to all their friends, family, and co-workers – new potential customers.