The insurance industry is preparing for a huge transformation in how it conducts business — and big data capabilities will lie at the centre of it
The insurance industry has been using data to price risk for years, but with technology now available to analyse large volumes of information for useful patterns, the value of understanding it has grown exponentially. Peter Littlejohns takes a look at five ways insurtech companies are using big data in the insurance process.
If it were possible to travel back 20 years and walk around the office of an insurance company, it’s likely it wouldn’t look too different to the way it does today — other than a lot less waste paper and overflowing ashtrays.
But behind the otherwise familiar veneer a vibrant discussion is taking place over the range of benefits that could be garnered through the right approach to big data — the popular name for large sets of numerical information that can be analysed to reveal patterns, trends, and associations.
Insurers are at an early stage in their journey with big data, but already multiple technology companies are vying for the spoils that come with helping extract financial value from it.
At the same time, some insurtechs see data analytics as their secret sauce and are keeping it under wraps from insurers, instead partnering with them to fuel their own business models with underwriting capacity.
Five ways insurtech companies are using big data in the insurance process
Pricing and underwriting
Estimating the price of an insurance policy based on complex risk assessment procedures is the bread and butter of insurance companies.
In its latest report, the European Insurance and Occupational Pensions Authority (EIOPA) found this was the area where big data is having the most significant impact on insurers.
The most obvious area in which this can be seen is motor insurance, where huge players such as IBM and LexisNexis Risk Solutions, as well as smaller companies including the likes of Swedish start-up Greater Than, are giving insurers the ability to more accurately price each policyholder by comparing individual driving behaviour with a larger pool of data — a process that allows them to correlate behaviour with risk.
But big data analysis also lends value to other lines of insurance, including life and health insurance, where patterns are established by associating behaviour with mortality and healthcare needs — often measured using wearable devices.
Due to the stark volume of data generated by insurers, big data analytics have only recently been made possible by advances in AI, specifically machine learning — a sub-discipline that requires the training of a computational model that can analyse large amounts of data in a matter of seconds at times.
A traditional claims journey is typically composed of insurers assessing the loss or damage to a policyholder or their possessions, which, at times, can involve a long and drawn out process where an adjuster recommends whether or not a claim is paid.
Although more beneficial for sectors with large amounts of claim data to draw upon — such as motor insurance — all major lines are benefiting in some way from analysing this information to better segment claims, and in some cases fully automate them.
While some organisations are opting to develop this capability in-house, others opt for an external product, provided by companies like MarkLogic and DataRobot.
Both of these technologies allow actuaries to delve into the vast amounts of data on claims and other variables by setting a machine learning algorithm to work on a specific question.
DataRobot adds the additional function of automated machine learning — technology that cleans up data and automatically selects the most accurate statistical model to use to answer each query.
By partnering with Robotic Process Automation provider UiPath, the insurtech can also execute and automate decisions related to big data, such as whether an insurer should pay a claim.
Gaining insight on healthcare of customers
Health insurance is big business in the US, and with open enrolment for 2020 around the corner — companies need to prioritise how they’ll stand out in the market.
One way they’re doing that is through drawing insights from big data in order to better recommend both immediate and preventative care to customers.
New York-based Oscar has been doing this for a few years now, connecting with medical record aggregators in different states to use patient histories to make permission-based predictions and recommendations to consumers on its platform.
According to CEO Mario Schlosser, combining medical history with real-time data gathered by Oscar’s telemedicine service — where users can chat with a physician and make prescription requests — has meant the insurer can intervene when patients are at risk.
One example he gives is a diabetic patient who ordered insulin and forgot to do so again before it ran out. One of Oscar’s nurses saw the mistake from the data gathered, travelled to meet the individual and took them to a primary care physician to ensure they didn’t suffer a diabetic episode.
Shaping policyholder behaviour
From the time that telematics was introduced to the insurance market, a key benefit has been shaping consumer behaviour in a way that eliminates risk.
Big data is the fuel behind this change, because it allows insurtech firms to see which policyholders are heading for a claim with their driving, security practices at home or even their healthcare (as mentioned above).
One company that has enjoyed success in this area is Hippo, a home insurance firm in the US that recently became a unicorn by reaching a valuation of $1bn.
The insurtech firm uses IoT devices it distributes among policyholders to monitor a range of activities in their homes.
By comparing the real-time data collected by the devices with wider information about household risk, it can intervene before a claim occurs by recommending a policyholder adjusts high-risk behaviour such as forgetting to lock doors or set alarms.
Improving customer experience through chatbots
Much to the chagrin of insurance agents, who have seen their teams cut down with the introduction of chatbots, AI is allowing insurers to more rapidly respond to customer queries by automating the response process.
The process of implementation involves training a machine learning model on a huge amount of data on policies, claims and other areas of the business — but the result is near-instant responses to customer questions.
It’s not only customers benefiting from chatbots either, as AXA Winterthur, the Swiss arm of the French insurance giant, recently launched the technology across 250 branches in its non-life cover department to assist sales agents and advisers.
In this case, the staff are able to query the chatbot, name SIRI-bution, to gain information needed to serve customers quickly.