At the heart of London-based MGA Aurora’s ambitions to “shake up” the commercial insurance market is a central vision – that of utilising algorithmic underwriting to provide speed, education and personalisation to brokers and customers, while optimising portfolio performance for insurers.
It’s a point of pride that Aurora has already managed to design and deliver that proposition, according to CTO and co-founder Bijal Patel (pictured left) who noted that the specialist health product live on its platform is algorithmically underwritten. Already several new products are lined up for launch over the coming months, she said, which will extend the complexity and capabilities of Aurora’s algorithmic framework.
But what is algorithmic underwriting? Going back to basics, Patel outlined that algorithmic underwriting is the automation of the manual day-to-day tasks that an underwriter would undertake to underwrite individual cases. Essentially, she said, it’s the automation of the risk assessment, underwriting and pricing processes using technology and granular data.
“This means that manual tasks that traditional underwriters would do at the individual case level are automated, meaning that they can focus their time on portfolio management and value adding tasks and are not involved at the individual case level,” she said. “Portfolios are managed by a real time monitoring of the portfolio, and analysis and calibration of these algorithmic models using machine learning and statistical techniques.
“This provides three key benefits in turn to the insurer. It reduces the expenses of writing the business because you need less underwriters, it ensures the consistent application of underwriting rules, and it improves the accuracy and granularity of pricing and underwriting for insurers.”
She highlighted that this also provides three key benefits to the other side of the insurance value chain - customers and brokers. It enables them to get personalised cover, more education about their risk profiles due to the level of data available and increased speed of access due to the real-time coverage made possible by algorithmic underwriting.
“So why now?” Patel asked. “Because more data is available than ever before. And technological advances in the market mean that the technology can be applied in new ways to streamline what underwriters do and the data that we’re able to access in real time.”
Contextualising what this means for the insurance market, Dorota Blaszczuk, head of algorithmic underwriting at Aurora, emphasised how through algorithmic underwriting, her team is able to better analyse trends and correlations in risk and better optimise its models over time. This allows for fast decision-making, she said, as well the quicker deployment of any decision made.
“This is all with the aim of being able to optimise portfolio performance,” she said. “The automation of manual tasks also enables consistency in decision-making and increases the speed of the quotation, as well as our ability to analyse a large amount of data very quickly. Overall, algorithmic underwriting is not just about automation but also about streamlining the already existing underwriting appetite and risk decision engine.
“So, while we are very excited about extending our framework of machine learning and more advanced algorithms, we are also always focused on getting the foundations right as well.”
The key differentiator to bear in mind about Aurora’s algorithmic capabilities, Patel said, is that it caters for the small and medium space, not just the micro or traditional e-trading space. As a result, its algorithms need to cater for a much wider range of complexity. Examining how algorithmic underwriting is opening up greater opportunities for customers in the SME space, Blaszczuk noted SME’s requirement of having broad yet tailor-made coverage.
“When we look at the statistics,” she said, “80% of the companies that suffer a disruption in operations close within 18 months. That’s basically because they’re not able to absorb the losses, as opposed to larger corporates which are able to keep much more risk on their balance sheets rather than insuring it externally.
“Even though there has been huge progress in the SME sector, at the larger end of SME, companies are still underserved. Again, looking at the statistics, 80% are still underinsured, and half of them say that their premium is too expensive. At Aurora, we’re trying to target and solve those problems.”
That mission starts with understanding why this coverage gap has emerged, she said, and a key driver that has emerged is the lack of a proven track record from these businesses. However, algorithmic underwriting uses a much wider range of data points, and is able to better evaluate the risks, allowing better prices for standard coverage from insurers.
Another problem is a lack of data, legacy tech and a lack of automated processes. This means that obtaining tailor-made, price-appropriate coverage is not just hard for SMEs, she said, but also very time consuming. Essentially, there’s a lack of appropriate products and coverage in the marketplace for these businesses at this time, which is where algorithmic underwriting can step up by allowing faster innovation and the faster deployment of new products.
“At Aurora, we have built an algorithmic underwriting framework, and we prioritise data enrichment, and we are automating the underwriting processes,” she said. “So, we are able to faster scale product development and innovation.
“I think it’s also worth noting that this brings benefits to the other side of the market. When we think about brokers, algorithmic underwriting reduces administrative burdens and allows more risks to be written. For insurers, this reduces expenses, and allows them to deploy more capacity. But the goal of this is making insurance for SMEs more affordable and more accessible.”
The insurance market generally has reacted with excitement to the opportunities presented by algorithmic underwriting, Patel said, but it has resonated especially strongly with Aurora’s broking network. On average, it takes about three weeks – and can sometimes take a couple of months – for brokers to get quotations from a range of insurers, she said, but now they can receive quotes in real time, as well as being able to tailor and flex cover to the customer.
“We’ve had a lot of traction with brokers but with our customer base as well, because customers in this market are looking for more digital solutions,” she said. “They want the same from buying insurance as they’re getting from other financial services but what they lack is the education. So, we’re prioritising our ability to use the wide range of data we have on customers’ risk profiles to educate the consumer and allow them to understand their risk profile and, in turn, how to mitigate that risk using insurance.”
What are your thoughts on this story? Feel free to share them in the comment box below.
