Ahead of the pack 

Prepare for augmented underwriting, says Lara Korz – or be left behind

Technology influences every part of our lives. As it continually develops and improves, it is important that we adapt to these improvements - and utilise them. The insurance industry's development in the digitalised era may have been a slow one, but technology is finally starting to influence our business practices.

One of the hottest topics of the moment is machine learning. It's a phrase that is bandied about all the time, but what does it really mean? In the most basic terms, machine learning involves using artificial intelligence to teach computers to think as humans would - in other words, using algorithms to parse data and learn from it. A computer is then able to make determinations and predictions on possible outcomes.

All about data

The insurance industry makes decisions based on an ever-increasing amount of data, which is key to staying relevant in the changing landscape. Underwriters who are able to extract value and yield higher accuracy using previously unrelated and untapped datasets will have an advantage and start to break away from the pack. 

The current uses for machine learning vary from chatbots to driver performance monitoring (telematics), fighting fraud and real-time pricing. However, there is still a long way to go before we will see the insurance industry effectively implementing and utilising co-operative artificial intelligence.

A combination of algorithms, such as a policy conversion model and a claims model, will provide two scores. The first is 'likelihood of conversion'. This on its own is not ideal, as not all business is good business. The second score is 'propensity to claim', which allows for identification of profitable business. Together, these scores provide an insight into the level of risk at the time of submission, allowing operational efficiencies. These scores also enable underwriters to adjust the risk selection or policy premium in order to tip the scales towards a more profitable position. 

The future is coming

'Augmented underwriting' is fast becoming a reality, and represents the future of insurance - but how do insurers implement it? 

First, it's important to make sure the core operating system is fit for purpose. There is little point in employing lots of highly trained data scientists if they end up spending the majority of their time acquiring, combining and cleansing data before they can harvest actionable insights from it. It is impossible for any business to operate in real-time if it is dependent on batch processing or delegated authority business coming in eight weeks after the risks have been bound.

The second key component of augmented underwriting is data enrichment. This will help address one of the main sticking points for firms - the need for an asymmetrical user experience (UX).

The end-insured wants minimum questions and hassle, but the underwriter needs maximum information to be able to price the risk accurately. Data enrichment makes the UX as slick as possible while giving the broker a treasure trove of data about each risk. As a result, it is no longer necessary to ask the client upwards of 55 questions before offering them a quote.

Once this is in place, brokers and insurers can use data science to help assess, grade and price risks. However, augmented underwriting certainly does not eliminate the need for human intervention, especially in the intermediated broker channel. 

Human contact is still necessary to ensure a deeper understanding of each client's needs and to handle negotiation, but technology prevents cognitive bias creeping in - it's there to guide the broker, suggesting when a risk is less likely to materialise and when the underwriter might want to increase the discount on price. Alternatively, it can suggest where there's a strong likelihood of a claim, allowing the underwriter to increase rates to cater for the increased risk.

Those who adopt augmented underwriting early will have an edge over their competitors, as they'll be able to use data to understand and mitigate risk more effectively. Configurable systems will also allow underwriters to manage exposure, as rates can be adjusted on a real-time, live basis. 

The advantages are clear - for both the big, established players, and for the niche classes of business that have previously felt too small to warrant the investment in technology. Augmented underwriting is the future of insurance, and early adopters will find themselves leading the pack.


Lara Korz is chief data officer Azur