All material subject to strictly enforced copyright laws. © 2020 Insider Engage is part of Euromoney Institutional Investor PLC.
Terms & Conditions | Privacy Policy | Modern Slavery Act | Cookies
Technology

A roadmap for analytics: Underwriting

The launch of an algorithm-only syndicate at Lloyd's speaks to many underwriters' darkest fears - that they could one day become obsolete

Abstract crowds of people with virtual reality street display
Credit:
gremlin/Getty Images

Analytics and underwriting - Key takeaways:

• Coronavirus has accelerated digitisation in the (re)insurance market, driving further disruption of traditional underwriting

• McKinsey suggests manual underwriting for personal lines and SME insurance will cease to exist by 2030

• In the specialty (re)insurance space there still be a strong need for human involvement in underwriting

• Analytics will help cut down the amount of data ‘noise’, to create insights at the point of underwriting

• The Cloud has democratised access to managed data services, dramatically bringing down cost and raising computing power

• Responsiveness to customers demands/needs will be the key differentiator of competitive performance

• New roles are emerging from the 'robot revolution', necessitating a need for re-skilling of underwriting businesses

• There is an urgent need for more data scientists and engineers, to identify the most useful insights to underwriters and claims managers

The launch of the Lloyd's market's first algorithmically-driven syndicate, Ki, in January 2021 has captured imaginations and heralds a new era in the use of data and analytics at the coalface of underwriting.

Launched by Brit in collaboration with Google Cloud, Ki has already received some serious endorsement, with backing from private equity giant Blackstone alongside Brit parent Fairfax.

The 'follow-only' syndicate will use a set of underwriting rules in code to deploy capacity. It is a "significant milestone" according to Brit CEO Matthew Wilson, who says Ki will "play a central role in the transformation of the specialist insurance market in London, creating a whole new and fully digital segment which operates seamlessly alongside the traditional marketplace".

On the surface it can be argued that now is definitely the right time for further disruption of the traditional underwriting process.

The Covid-19 crisis has accelerated many of the ongoing digitisation trends and initiatives within the market, with data from the London Market Group (LMG) revealing that usage of PPL swelled during lockdown.

Meanwhile, Lloyd's has increased its targets for e-placement submissions for the second half of 2020.

"Since March, the working world has changed significantly, with [people in] the industry unable to physically see one another and conduct business like in years past," says John Merchant, managing director, cyber - US and Canada at Optio.

"With the arrangement of face-to-face meetings becoming extremely challenging, the advancement of carrier, broker and wholesale producer platforms and investment in pushing information out by APIs has accelerated."

Art versus science

On the other hand, is the market really ready for underwriting via algorithm? It is taking the age-old 'art or science' debate to a new level, at a time when clunky legacy systems are still holding many market participants back, along with a lack of investment in data collection and curation.

Perhaps more crucially, how does the launch of a venture such as Ki allay fears within the underwriting community that big data and advanced analytics are friends, not foes?

"For any new technology there is this underlying concern that 'I may end up being obsolete’," says Merchant. "And that's where you get pushback from the underwriting group."

According to consultancy firm McKinsey, within personal lines and SME insurance, manual underwriting processes will cease to exist by 2030. In this commoditised space, it believes the process of underwriting will be reduced to mere seconds as decisions are automated, supported by a combination of machine learning models, and powered by a deep pool of internal and external data, accessed through APIs from data and analytics providers.

However, it is a different story within specialty classes of business, thinks Merchant.

"I do believe when you have an underwriter doing personal lines or something heavily-automated they're no longer really underwriting...they're actually sales people," he says. "But in the specialty space there is still a tremendous need for underwriters to be involved."

Cutting through the noise

At the same time, underwriters must accept that data and analytics are not going away and use the tools available to them to differentiate themselves.

The drive for efficiency is important, as are commercial pressures to grow the business, but the real battleground will be the ability to distil data down to actual insights at the point of underwriting, in order to optimise (re)insurance books of business.

"One of the issues underwriters have with cyber risk is they say there's not enough data," says Merchant. "There may be too much data actually. The problem is there's a tonne of noise in the data and we’re not certain what actually correlates with the loss. Curating that data will allow us to gain a level of certainty from that noise - to see a signal."

"If data suggests a negative outcome, underwriters must take that into consideration and show some discipline," he adds. "It's about using big data and analytics to avoid bad outcomes, as opposed to digging one’s way out after the market turns sour."

The fact that the processing power to carry out advanced analytics is now readily available has already levelled the playing field, it is argued.

"Data was always there, but the volumes are now much bigger, and our ability to handle and process it has gotten better and faster," says Bart Patrick, managing director for Europe at Duck Creek Technologies.

"A key change has been in managed services, which have been around as a concept for some time, but now cloud [computing] has democratised this, dramatically bringing down the cost and incrementally raising access to almost limitless computing power. This processing power has changed the application of data," he says.

Customer outcomes will become the key differentiator, thinks Marcus Broome, chief platform officer of Whitespace.

"Suppose the actual end result of the insurance would be the same price, however someone has arrived at it? It's about how, in those circumstances, do you compete and always win the business."

"All of us - in any transaction - tend to buy normally from the organisation that's most immediately responsive to our needs," he continues. "What this is enabling is a shift in response to the customer. We believe that anybody working with a digital contract will always outperform someone competing using traditional technology because they will be getting back to the customer immediately with a response."

Time to reskill?

Fears that jobs will be taken over by robots and computers are far from unique to the (re)insurance industry. A report by the World Economic Forum (WEF) has found the global workforce is automating faster than expected and will displace an estimated 85 million jobs over the next five years.

"Covid-19 has accelerated the arrival of the future of work," according to Saadia Zahidi, WEF managing director. "The window of opportunity for proactive management of this change is closing fast."

Yet new jobs are emerging from what the WEF calls 'the robot revolution' and this brings with it an urgent need to re-skill.

By 2025, analytical thinking, creativity and flexibility will have become the attributes most needed by the modern workforce, with a surge in demand for talent at the forefront of the data and AI economy. Such skills will be essential for the next generation of underwriting professionals.

There is also an urgent need for more data scientists and engineers, to identify which insights are most useful to underwriters and claims managers.

"There's got to be a recognition that these tools can help tremendously," says Optio's Merchant. "Because when you're set up with data and analytics, after the initial cost, the scalability of them means the upkeep is much less than hiring more and more underwriters."

"There are costs associated with people that are not associated with technology. So we need to get some trust in these tools as ones that will help us as underwriters."

Share
We use cookies to provide a personalized site experience.
By continuing to use & browse the site you agree to our Privacy Policy.
I agree