Assessing risk on the road to full autonomy
With the proven potential to improve road safety and save lives, Advanced Driver Assistance Systems (ADAS) have undoubtable benefits for both the insurance industry and society in general.
Yet there are also a number of challenges to consider as the adoption of these systems become more widespread. How can insurers harness the data that they offer up and how will this change the profile of motor risk as we know it? How can we ensure that motor insurance products stay ahead of the curve and offer consumers a fair price without factoring in a margin for the 'unknown'?
Put simply, ADAS perform a number of tasks that enrich the lives of drivers. Think of driving now without cruise control or the reassurance of tire pressure monitoring. The very necessity of these systems for drivers is why the global market for ADAS is expected to reach more than $67bn by 2025, growing more than 10% each year. A group of 20 carmakers has pledged to outfit almost every new vehicle with forward collision warning and city-speed automatic emergency braking by 2020 (Reuters, 2019).
There are also signs, however, that the adoption of ADAS is unequal across different demographics. The UK government recently concluded that ADAS could improve the lives of older drivers by increasing safety, helping keep on top of maintenance and, ultimately, keep those drivers driving for longer. In turn, it could help offset the negative consequences of poor mobility in older age including isolation and depression.
Clearly, the potential advantages for this group alone cannot be overstated, yet there are also signs of unequal adoption between different age groups and genders more generally.
Creating motor insurance products that are more cost effective for consumers could be the key to addressing this disparity.
Are we there for full autonomy?
There is still a lot of ground to cover on the road to full autonomy and there have certainly been challenges along the way.
As a pioneer of several such autonomous technologies, Tesla, for example, has made headlines on a number of occasions due to accidents involving cars with ‘Autopilot’ activated; and the UK government recently reminded drivers that the feature is not for legal use on UK roads.
Nevertheless, there is significant evidence to show that certain ADAS features are improving safety overall.
Research we conducted in conjunction with the BMW Group revealed that Emergency Brake Assist alone has the potential to reduce rear-end accident frequencies by more than 30%.
Likewise, other basic ADAS technologies like forward collision or lane departure warnings could reduce accidents by 16.3% on motorways and 11.6% on other roads. More sophisticated ADAS technology like AEB and lane-keeping assistance could have an even greater impact, reducing accidents by 25.7% on motorways and 27.5% on other roads.
Calculating premiums to meet the needs of ADAS
In this way, the rapid introduction of ADAS is proving its potential to keep drivers safer on the road, but in order to achieve this, the technology needs to be managed correctly and safely.
As these technologies become more prevalent across the UK car market, the process of calculating insurance premiums becomes more complicated. Whilst the correct use and maintenance of these systems by an accredited technician should result in fewer insurance claims, the knock-on effect is a rise in claims costs for insurers, as repair work requires more complex technology.
To remedy this situation, insurers must work closely with motor manufacturers to secure the higher level of detail they need on ADAS technology, allowing them to price car insurance premiums more accurately.
To help tackle this problem, we have created our own proprietary ADAS Risk Score, which uses hist oric datasets to compare the accident frequencies and claims severities of individual vehicles with and without certain ADAS technologies.
We also conduct advanced simulations, which enhance the score by providing greater insight on the most frequently claimed incidents.
Finally, we perform live tests of ADAS technologies in a variety of real conditions, such as testing the effectiveness of AEBs and other ADAS on the test track.
Contextual data as the next frontier for insurers
Notwithstanding the emergence of new risks around data and software security within the car, we expect that the growth of car connectivity and ADAS in the long run will reduce the overall risk and the frequency of accidents.
The big opportunity for insurers lies in harnessing data from vehicles and other sources to more accurately price the risks of insuring drivers. How rapidly data-driven models can be deployed of course depends on our ability to capture, process and harness that data.
We also observe the abundance of data sources and the rush to monetise it as quickly as possible. Efforts are underway to change this fragmented picture. As a first critical step, the automotive industry is already working to agree on a standard way for vehicle sensor data to be transmitted to the cloud for aggregation and analysis.
This will enable the entire industry to benefit from the processing of data at scale and to create more accurate and precise traffic services and road hazard warning systems.
In the meantime, newer entrants to the vehicle technology market, such as Tesla and Daimler, are already having much success. Usage based insurance models will continue to improve and perhaps become more prevalent both in the personal and commercial vehicle space. And we will continuously expand our understanding of the most recent vehicle technologies.
Some are moving faster than others in building the capabilities needed to assess these automated features and to develop data-driven approaches to risk calculation. But without a doubt, going forward, differentiation and embedding new technologies will be key for insurers to prosper amid this change.
Insurtechs are winning the race with legacy system companies
Nestled in its own place within the world of financial services, insurance is arguably more unpopular than retail banking.
That’s hardly surprising given that, from a customer service perspective, insurance is something of an off-kilter transaction. You pay a sizable premium in exchange for a service you hope you will never have to use. This image problem is exacerbated by ubiquitous tales of insurers not paying out when it is time to make a claim.
The insurance sector has long been due to an overhaul, and this is where the disruptive force of insurtech comes in - one of fintech’s most upwardly mobile subcategories. Accordingly, last year, insurtech in the UK alone attracted £262m in investment, a growth of 60% on 2019, according to Tech Nation. Insurtech’s momentous growth has been captured in a new report by The AI Journal exploring this burgeoning sector.
What exactly is insurtech?
Put simply, insurtech refers to technological innovations that seek to make insurance cheaper to buy and more efficient to use. In a similar vein to fintech, the large, established institutions have been dipping their toes into insurtech, but it’s the disruptors who are genuinely looking to shake up the status quo, diving into and exploiting those areas that traditionalists have little imperative to explore.
Examples are price comparison sites (one of the earliest forms of insurtech that was eventually snapped up by the insurers it initially sought to disrupt), claims software, customisable policies, or even smart-tech-enabled dynamic policies whose premiums can fluctuate depending on changing circumstances.
The latter, for instance, could use someone’s fitness tracker or smartwatch to monitor fitness levels, thus reducing the premium of a life insurance policy; or track a GPS system that records the location of a car and assesses risk levels accordingly.
Most consumers tend to shop around for their insurance needs and perhaps end up buying their contents insurance with one provider, their car insurance with someone else, and their pet insurance with yet another underwriter. Managing all these different policies, with their varying renewal dates and payment terms can be complex. This has led to the increase in apps that pull everything together.
More prosaically, insurtechs are developing AI that uses machine learning to act as an insurance broker, eliminating the need for a human intermediary and therefore offering more cost-effective and impartial advice.
Insurtechs and risk
But there are some obstacles in the way of insurtech’s continued evolution.
Insurance companies are averse to risk. Understandably so, as at the crux of the industry is the role of the actuary, whose job it is to analyse and measure the probability and risk of future events. So it’s little wonder that there’s a reluctance among the traditional players to welcome the disruption that insurtech brings.
Insurance is heavily regulated, a minefield of legality and labyrinthine jurisdiction, which means the idea of shaking it up can be anathema. And why would they, when their old-school business models are working perfectly fine?
There’s an understandable nervousness and unwillingness to work with startups, who themselves need to work with the bigger firms in order to underwrite risk.
While it seems like a catch-22 situation, there is growing, if cautious, interest from insurance companies, who can see the benefits of insurance with a friendlier face, innovative solutions, and a competitive edge through differentiation. As that tentativeness dissipates, the growth of insurtech will gather even more momentum.
Tom Allen's analysis is based on the findings of a new report on the fintech and insurtech industries produced by The AI Journal.