Optimise existing models effectively, and understand how to leverage new ones

In a period of intense change, it is tempting to plunge wholesale into new technologies and philosophies. They seem to promise so much. But it is important to recognise the valuable assets that already exist within an organisation and treat innovations as the opportunity to augment those capabilities.

In a session at Underwriting Innovation USA on Tuesday November 9, Rangarajan Srinivasan, head of insurance solutions at HCL Technologies, will be demonstrating how attendees can meet new challenges and make sense of innovations in the sector. He’ll reveal the effective processes and technologies that can make the most of artificial intelligence (AI) while still maintaining smooth, day-to-day operations.

Srinivasan spoke to Intelligent Insurer ahead of the event.

What is troubling the carrier of today?

Carriers face three broad challenges. First, as a legacy business, a lot of the inputs are unstructured—be they snail-mail documents or spreadsheets. Many carriers are still scanning and indexing. There may well be digital versions of documents but the underwriters will still be typing or looking up information.

The second is that many insurers aren’t using the data they have internally, and they’re sitting on a diamond mine. There is an explosion in third-party data that can provide insight and deeper analytics. All of this has to be brought into an enterprise data strategy. But to be able to use it even at a simple level, you still need to construct it.

Finally, there is the legacy framework. Carriers may well have access to the latest analytics but they’re still running software that is 40 years old, so they can’t integrate it into the business. Even if they have a suboptimal solution for data and unstructured inputs, the whole underwriting process is so inflexible as to make it almost impossible to react in a more agile, modern way.

How do carriers meet these challenges?

What we are proposing is an exact counterpoint to those three challenges. First, it has to be an ecosystem. No-one will say: “I will be everything you need.” Some say it’s about selecting the best of the breed, but the neatest way to think about it is as a Lego block.

You could even think about a single-point solution: when you stitch it all together you can see the end-to-end process. The point is to create micro-projects that can start plugging holes which can then connect to create a larger transformation.

There are insurtechs with solutions that will do the good stuff with your documents and give you a set of application programming interfaces that can plug straight into your system. There are data solutions that can provide you with insights. But one of my underwriting clients told me that they don’t know what will be useful.

We put the same effort into every submission that comes in. The chances are that underwriters are putting a lot of effort into a potential client who will not ultimately convert.

How can carriers use technology to triage their risk appetite?

There needs to be a data solution that can help understand the propensity of a potential line of business to be good, bad or ugly. Then carriers can, at least, make the call: they can ignore the ugly and perhaps, at least, look at the bad.

A ton of things are available today for solving a whole range of issues. The best way to tackle this is with a phased approach that doesn’t break the bank or make a project too large to deal with.

How do carriers begin to select the right technologies and create a ‘toolkit’?

Working in an ecosystem means carriers can use external data and advanced analytics to get much better insights from customers around underlying risk. Then, they will be in a better position to create more prescriptive products for their clients. They can say: “I know you want to get coverage ‘A’, but you may benefit from the core of ‘A’, plus ‘B’ plus ‘C’.”

As the ecosystem gets better, you can move into creating predictive systems. One of the best data alignments is the internet of things (IoT) ecosystem, where you can find more advanced real-time information about an underlying asset or property—whatever is being covered—then do predictive loss mitigation and make comparisons.

For example, if 100 customers are covered and 50 have some IoT infrastructure, they become the people with fewer losses and, therefore, lower premiums. That allows you to pitch an IoT product to those who don’t already have it. It’s not about selling more but selling non-insurance services enabled by the partner ecosystem.

It might sound futuristic, but there are chief information officers/risk officers engaging with this today.

What would you like attendees to take away from your session?

AI and machine learning (ML) both provide rich risktech tools but underwriters also need to know what they can do today, without necessarily jumping onto the fancy stuff. We can price, show five or six parameters and create scenarios for underwriters, just by using historical data.

Play around with these things and see if they can be optimised. It doesn’t necessarily mean a premium will immediately go up or down, but it is a very practical solution. And then, six months down the line, there will be a raft of data from underwriting decisions across multiple different scenarios that can be used to implement an AI/ML model to deliver profiles. It’s about reaching those AI- and ML-driven solutions in a logical way later, rather than trying something big now.

Rangarajan Srinivasan, head of insurance solutions at HCL Technologies, will be demonstrating how attendees can meet new challenges and make sense of innovations in the sector during a session at Underwriting Innovation USA on November 9, 2021, at 12:55 pm EST.

This is the biggest online event tailored specifically to underwriting—don’t miss out. Register now!