Spotlight: Motor trends - the use of data in claims and underwriting

Over the past two years, every sector has experienced an operational environment that can be described as ‘atypical’; and insurance is no different.

A recent survey conducted by Post in association with Verisk identified key themes – and some particularly interesting trends – about more structural changes occurring within the motor claims and underwriting processes.

Two primary concerns that respondents articulated were how more external
data can be implemented into their business for both underwriting  and claims. We’ll take these two areas in turn, beginning with claims.

Part One

Resistance is futile. The future of claims is digital – and it’s coming soon

While Covid-19 has had a massive impact on the insurance industry overall, the motor space saw a massive shift in the 2021 survey, due to so few people using their cars as usual over the previous 12 months.

As a result, 88.5% of claims respondents said there were fewer cars on the road and fewer miles driven, resulting in an 84.6% reduction in accidents (see figure one). However, when there were incidents people resorted to different channels to report them.

Aviva saw its online claims spike by 50% in March and April 2020, though it did later drop. However, Aviva has, what it considers, high levels of digital reporting, with up to 70% of all motor claims for direct customers now being reported online, said Neil Ashley, head of digital claims at Aviva.

Customers can’t be divided by demographics as easily as they might once have been, Ashley continued: “It’s a broad range of customers that use our digital services. Though we do find that customers who have purchased their policy online are more likely to claim online.”

Although accidents were down, 53.8% saw the cost of settling claims increase. Remote working may have made settling claims more difficult, suggesting there is a lack of automation in the claims process. Only 15.4% saw the costs of claims reduce in this period.

Rapid adoption of technology

Unsurprisingly, self-service for first notification of loss is being used by the vast majority of companies surveyed. More than two thirds of insurers are using it, with most of the rest looking to introduce it in the future (see figure two).

Video loss inspections are used by more than three quarters (76%) of companies with telematics close behind on 73.1%.

Though insurers clearly have digital engagement in terms of policyholder portals for this self-service, there isn’t a great deal of joined up technology.

It’s interesting to note that, while AI is deployed in more than half of the companies surveyed, it has not been extended beyond estimating or simple interactions in the claims process, in the form of chatbots. Almost half (46.2%) identified AI claim processing as a consideration for future use and appointment scheduling systems, 42.3%.

“AI can make more accurate or quick decisions as to whether a vehicle may be repairable for instance, removing delays for the customer,” said Aviva’s Ashley.

This is also true of appointment scheduling, which can allow a customer to go through an automatic cover check automatic when reporting online as to whether it can be repaired and provide that customer with a link to book their repair with a garage in the network.

“We are already doing that,” added Ashley, “and that’s a completely touchless journey.”

The Area identified with the greatest growth was AI claim processing, which is currently only used by 30.8% of companies. Appointment scheduling systems is the next area primed for growth and is currently only used by 23.1%. 

However, not everyone will have such a seamless process and those who can’t necessarily compete will game the system.

Someone Post spoke to said that sometimes people can be “hoodwinked” into thinking a digital process is going to customer self-service, but doesn’t when it is full of holes behind the scenes. This results in a lot of customer dissatisfaction.

“There are plenty of solutions that appear that allow you to put yourself into a body shop nearby,” they added. “But in reality, what [truly] happens is the body shops make all [of] their slots available, because they never want to say no to a customer, even if they don’t [actually] have the space.”

More digital, please…

Yet, there’s considerable optimism about the increase of automation within claims. The survey showed that within five years, that almost half (46.2%) thought 41-60% of claims could be fully automated without any human intervention (see figure three).

Nevertheless, while processes are important for operations, digital engagement remains a key objective for insurers. None felt it to be unimportant or of little importance, with all considering it important and expected by most policyholders (57.7%), or critical and demanded by the majority of policyholders in 42.3% of cases (see figure four).

The pandemic made digital channels essential for business to be conducted. But insurers have also been told through customer feedback – and behaviour from those who do not interact with telephone processes – that there is a need for another channel and digital is the only sensible option.

Peter Simon, chief trading officer, at Esure said: “Operationally we can use insight from AI gathered data to better understand our partnerships, which means we can offer more integrated and stress-free experiences. Our new eFNOL journey gives customers the ability to initiate a claim quickly, using an innovative, conversational interface, utilising a chat bot as the first point of call, delivering a simple, intuitive and efficient experience.”

Part Two

Digital technology is already transforming the underwriting process

The biggest challenge underwriting faces in insurance companies is the constraints and limitations set by existing legacy technology. This came as the top of the league of obstacles and received twice the weighting of fraud which was in second place (see figure five). Quality of third-party data sources was some way behind fraud in third place before renewal retention, access to analytics resources and deficiencies in vehicle data.

Half of respondents (50%) said that customers had increased demand for digital engagement. The data may not show it, but it remains very possible that had there been greater provision of digital engagement, and some of those increased costs may have been reduced during the Covid-19 pandemic.

“AI is becoming central to the insurance industry, especially considering the volumes of data we need to assimilate, understand and learn from,” said Simon.

This requires an adequate data strategy to be run alongside sufficient technology and infrastructure to support the use of AI, continued Simon: “AI isn’t just about predictive modelling. It’s about a company mindset, a cultural change, and a unified approach across the business to understand its needs.”

Simon claimed AI and chatbots can be used effectively to provide customers with an immediate and responsive first point of contact, yet you cannot take away the human touch completely if the company needs or requires it.

It’s what you know that counts

There is an interesting shift underway in the use of external data sources for underwriting over this period. ABI risk groups (67.6%), location-based data (62.2%), vehicle history and Motor Insurance Anti-Fraud and Theft Register data – both 45.9% – remain the most commonly used forms of data (see figure six). But in future, telematics, advanced driver assistance systems accurate mileage data and open banking data and are going to be the key targets for respondents to the survey.

“Telematics and ADAS is one of those fundamental initiatives that will allow us to think about rating risk in a slightly different way,” said Nimesh Patel, CEO, Wrisk.

“More data signals are better and yet the primary movers of what determines your price tends to be your age, credit score, your postcode, as it always has been.”

The potential paradigm shift will come with the capture of the safety features embedded within the car model, for instance, as it may be more or less safe than another vehicle.

Then it can become more personalised by taking account of driver behaviour. Mileage is one of the first things that determines price, but the insurer has very little view of verifying of ow or when the customer drives their 5000 miles.

Usage-based insurance concepts capturing yield type data on driver behaviour clearly has a massive opportunity to process that data at scale, to leverage the concepts of machine learning, and be able to start rating risk more accurately,” says Patel.

There’s a treasure trove of data available within the vehicles, and if insurers can get it directly from the car manufacturers, there will be a major step change, says Patel. If autonomous and semi-autonomous vehicles make driving safer, premiums might well go down at some point in the future, so there’s a lot more to think about than just the additional protections for range anxiety, cable cover to capturing data from when the car is in motion.

Crunching the numbers

Currently, the location based data, ABI vehicle history and then consumer consented data are currently considered the most important. This may change with the adoption of these other technology-based data sources.

When it comes to using external data sources, there’s no lack of decent coverage or data quality itself, or in-house analytics capability according to the respondents. The main obstacle is how to interpret and use the data, having time to identify suitable data and then, how to make use of it (see figure seven).

After this is the lack of IT capability that makes implementation hard. This is the perennial chicken and egg situation faced by the insurance industry. It has been suggested that digital transformation will in fact increase the ability of insurers to interpret data.

The need for traditional actuarial roles may be reduced by technology, but this means that this valuable resource can be released to provide greater input on the interpreting the data for the purposes of underwriting and product pricing.

Of course, the systems need to be capable of analysis and many are not even close to that situation, yet. But it is coming and these top three challenges are completely interdependent. Despite the appetite for bringing in telematics and driver behaviour data, almost a third (30.56%) believe it’s moderately unlikely that these new data will replace current data used for underwriting risk models. However, almost one third (30.56%) believe that these very likely be the default data categories used when assessing medium to high risk cases. Nevertheless, fewer than 10% are satisfied they’ve got enough data following the most recent insurance pricing reforms. Exactly a third are unsure, while almost 60% are looking for more data to support their pricing decisions (see figure eight).

Tailoring cover for the individual

There is also a need to understand how useful data that can now be accessed will improve processes.

Geoff Carter, CEO of Sabre, believes ADAS features will become increasingly important in rating, and is already feeding this into its rating models. But there needs to be more caution around mileage, he warns.

“To state the obvious, a car isn’t driving anywhere if it’s stolen – and that’s a substantial cost. The context of miles driven is also important – a vehicle being driven at sensible speeds on a motorway may be far lower risk than one doing fewer miles in a town centre surrounded by children and other vulnerable individuals.

“As a business we have always been wary of telematics as the discount customers required for being monitored seemed disproportionate to the decrease in risk.

“If telematics can transition into more of a value added model, or becomes embedded in car technology, this may change.” 

Callum Rimmer, co-founder of By Miles is convinced that this is exactly what is likely to happen in order to improve underwriting. By looking at usage – when and where customers are driving, the routes they taking, etc – the customer can be encouraged to change their behaviour.

“You can benefit insurer from a loss ratio point of view, but you can also encourage the customer to do it by charging less premium if they use their car less or differently.”

New world order

Telematics, or data on a particular customer is really the best tool to ensure you’re giving the right discounts – ie renewal premium – to the drivers who deserve it,” says Mike Brockman, CEO of ThingCo.

While telematics has been adopted by the UK motor sector, it has been rather half-hearted and in the past year or so, has been flat, says Brockman. He’s convinced that situation is about to change.

“Demand should go up this year, as insurers using telematics will suddenly become more competitive against non-telematic competition, with the changes introduced by the regulator on renewals,” says Brockman.

The arrival of autonomous and semi-autonomous driving will also deliver “another paradigm in underwriting”, says Patel.

“The development of connected and autonomous vehicle will completely change the underwriting of car insurance over the next decade,” he adds. 

Sponsored