Original article was published on Deep Learning on Medium
The term “insurance”, in a purely legal exegesis, refers to that legal transaction with which one “guarantees” oneself from a future and uncertain event.
A rather obvious corollary of this is that the insurance sector is based on data: it is the data, collected and analysed on a case-by-case basis, that guides and qualifies the risks and, consequently, the costs.
In a context, like the present one, in which we look at the involvement of new technologies with a mixture of hope, in a significant contribution, and fear, of being “replaced”, the use of Artificial Intelligence can really make the difference!
Artificial Intelligence encompasses many technologies, whose future use will have a significant impact on a predictive market:
· there is machine learning, used to identify, in data, models, and in this way make predictions;
· there is deep learning, application of the previous one but able to act autonomously, because the model learns from data and, without instructions, draws conclusions;
· there are neural networks, which imitate the human brain, from data, elaborate models, even complex ones, without the need for human intervention;
· and there are chatbots, which imitate natural language, understanding and formulating responses in written or vocal form.
McKinsey estimates a potential value of up to $1.1 trillion per year if AI technology is fully applied to the insurance industry.
Andy Cassel, vice president of data and analysis for insurance software, points out that the introduction of AI in an organizational context requires not only technological or procedural change, but global change. «No one has to use AI to do their job well,» he said, but «What you have to do is make AI a natural part of the work that makes sense and offers value to those who use it».
The step is not immediate, it requires first of all the digitization of all that data, mentioned above, which are now still “imprisoned” in paper documents, called “dirty data”.
The result would therefore be a customizable and competitive service, tailored to each customer, who would only have to pay for what they actually need. Then there would be an improvement in marketing and commercial aspects. In addition, even the aspect of pure assistance, or customer satisfaction, could benefit from the use of chatbot, in aid of the insurance staff. And that’s not all! Also the risk of possible frauds, through neural recognition systems, would be significantly mitigated, leading to more competitive premiums for all and lower policies.
This would result in a strategic improvement in health insurance, where customers could pay for the services they need (pay-as-you-go), and customize their needs; in motor insurance, by greatly improving the service, for example by photographing vehicle damage in real time for a claim and already having an assessment of repair costs, handling claims in six minutes and 48 seconds, as would the insurance company, created by Alibaba, Ding Sun Bao.
Artificial Intelligence can significantly improve production processes and in some areas more than others, insurance is one of them!
All Rights Reserved
Raffaella Aghemo, Lawyer