X

Log in

Forgot password?

Don't have an account with us?

Register Now
instech-london-podcast-150-lloyds-event-website

Ed Gaze & Colin Masson

Lloyd’s Lab & Beazley

Complex claims innovation

Claims is a vital component in the insurance value chain, but despite its importance it doesn’t always get the innovation attention and headlines that it deserves.

On this week’s podcast, Matthew is joined by Ed Gaze from the Lloyd’s Lab and Colin Masson, Head of First Party Complex Claims at Beazley, and member of the Future at Lloyd’s Claims Design Group, to discuss innovation and new technology solutions for settling claims for complex risks.

Episode 150 also features insights from Tautona, Safekeep, SettleIndex and distriBind – four companies who have recently built solutions in the Lloyd’s Lab to improve claims processes and the customer experience.

Talking points include:

  • The role of the Future at Lloyd’s Claims Design Group
  • The unique challenges of settling claims for complex risks
  • The fundamental elements of claims innovation
  • How data, analytics and AI are improving claims processes
  • The goals and benefits of the Lloyd’s Lab

The conversation was recorded from the InsTech London Live Chat, The Future of Lloyd’s and Claims Solutions for Complex Risks, which is available to view on our BrightTALK channel.

If you like what you’re hearing, please leave us a review on whichever platform you use, or contact Matthew Grant on LinkedIn

Sign up to our newsletter for a fresh view on the world every Wednesday morning.

Continuing Professional Development – Learning Objectives

InsTech London is accredited by The Chartered Insurance Institute (CII). By listening to an InsTech London podcast, or reading the accompanying transcript, you can claim up to 0.5 CPD hours towards the CII member CPD scheme.

The Learning Objectives for this podcast are:

  • Discern some of the important elements of innovation
  • Realise how technology companies can integrate innovation into existing workflows
  • Understand some of the things that are holding innovation back

Let us know you have listened to this podcast and if your organisation is a member of InsTech London you will receive a quarterly summary of the CPD hours you have earned.

To help us measure the impact of the learning, we would be grateful if you would take a minute to complete a quick feedback survey.

Complex claims innovation – Episode 150 highlights

Ed Gaze, Senior Manager, Lloyd’s Lab

For those who don’t know about it, what is the Lloyd’s Lab?

Lloyd’s is a historic and interesting place where specialty risks get insured. It’s not an insurance company – it’s a market of over 50 insurers. The Lloyd’s Lab is at the heart of it, and it’s great because companies that are working in the Lab have access to 50 potential customers. 

What’s the goal of the Lab and what benefits are there to companies that apply?

The purpose of the Lab is to bring companies with new innovations that can support insurance and insurtechs from around the world into the Lloyd’s market. This allows these companies to experiment with the market, try new ideas and ultimately contribute to the modernisation of Lloyd’s. We want companies that are creating new insurance products or making the existing way we work more efficient. If companies think they’re doing that, we want to hear from them. We don’t charge any money for coming onto the Lab.

How does the Lloyd’s Lab selection process work and how are cohort companies chosen?

We have four key steps: scoring, longlist, shortlist and then pitch day. Applicants are scored by people in the Lab and in the market, and from this the longlist is created. We then have a series of workshops that we use to shortlist for the pitch day. This is when companies get to present a short pitch and answer questions, and the market is there to vote again – it’s their votes that determine which companies make it in. For the last cohort, 117 companies applied, 60 were longlisted and 24 were shortlisted to pitch. 

The companies in cohort 6 are solving real problems that the market is facing. We sometimes see companies apply saying that they can solve everything, but that doesn’t solve the specific problem that the audience wants. There’s a big audience and a lot of problems, so companies should choose something specific that they can solve. For claims teams, it’s all about proving that they understand the problems that Lloyd’s and the market is facing, understand the customer and that they have a solution. 

How do you engage brokers?

With the way the Lab is currently structured, all mentors have to be managing agents. This year we are looking at how we can bring brokers into the mentor group. We often end up reaching out to brokers – we have innovation advisors that are brokers, so we engage them in that way. We’re thinking about how we could potentially make this more formal this year.

Colin Masson, Head of First Party Complex Claims, Beazley

What does the Future at Lloyd’s Claims Design Group do?

The Claims Design Group is a group of claims professionals from across the market, both Lloyd’s and company market carriers, as well as brokers. We’re supporting the initiatives described in Lloyd’s Blueprint One and now Blueprint Two with a focus on complex claims. We’ve been testing the ideas put to us and are carrying out strategic thinking and figuring out how we turn the ideas into reality. It is easier said than done – it requires a lot of commitment and effort from across the market.

What are some of the challenges that the group is trying to overcome?

Within the Lloyd’s market, over 50 managing agents are concentrating on different things. Lloyd’s is a place to underwrite the most complex risk in the world, right the way through to the smallest risks. This variety creates the complexity of how to manage it all on a claims  platform that’s efficient and can get everything done. With the subscription market, it’s often not just one person making a decision, so information has to flow to the right people quite quickly. 

What are examples of the types of complex claims you look at?

Although it’s exciting to see that windscreen claims can be automated, for example, that doesn’t provide a solution to the kind of things that we’re looking at. TComplex claims include for example hurricane loss on a Caribbean island or a kidnap and ransom claim on a vessel off the coast of Nigeria – within the Lloyd’s market there are a variety of issues. There are claims where the decision making is simple so they can be easily automated, but with the rest we need the right technology and information. When dealing with complex claims, having a package of information that includes access to information such as who the experts are and if a premium has been paid can make the process a lot smoother.

Is there anything that you’re seeing around data or analytics that will be a part of your ongoing innovation?

There are three fundamental things that matter when it comes to claims: people, technology and data. These all need to be best in breed in order to get the optimal outcome. In the Lloyd’s market, we have struggled with making sure that data is packaged in a way that allows us to feed it through the system. Without looking at how data gets inputted and how it’s sourced and enriched, we can’t get very far. We need to start with making sure that data sources are the best that they can be.

John Holdsworth, CEO, Tautona

What does Tautona do in relation to claims?

Tautona uses artificial intelligence to automate tasks that have historically required human judgement. Automating simple policy terms and conditions is relatively straightforward, but deciding coverage for a complex claim is a much more substantial task. It involves understanding what caused the loss, applying proximate cause doctrine, what was damaged or who was injured and if there’s coverage for the loss. Property insurance policies consist of multiple policy forms with hundreds of clauses and combinations, depending on the circumstances of the claim. Through working with Atrium and Lloyd’s, Tautona has converted all of those physical documents into a machine readable format using a combination of machine learning and natural language processing (NLP). This means that we can analyse policy wording in real time and determine coverage with limited human involvement.

How have you automated claims documentation?

No claims automation company can automate 100% of claims – a percentage of claims are always going to be referred for human assessment, usually somewhere between 20 and 25%. What companies want is to be able to triage the simpler claims, and automate them end-to-end using straightforward processing – these make up 75 to 80% of claims.

Can you talk about some of Tautona’s machine learning and AI?

We’ve developed an AI-powered claims automation system for the property line of business that acts like an experienced insurance professional. For standard claims, the application handles the process from first notice of loss (FNOL) all the way through to property inspection, with minimal human involvement. It also assigns claims it’s not authorised to settle or ones where a concern is identified. Even when human input is needed, our claims adjuster bot will have already done a lot of the heavy lifting. By harnessing AI at the beginning of the claims process we accelerate every part of it, creating efficiencies for the insurer, managing agent or the TPA, and allowing for greater clarity for the policyholder. 

How easy is it for a company to integrate Tautona into its current workflow?

There are a few carriers globally that aren’t ready to fully automate the claims process. We provide a plug and play approach – we’re not looking to replace a company’s claims automation solution. What we want to do is to add value to a company’s existing investments. Our specialty is around automating policy coverage analysis, specifically for property claims.

What advice would you give to those thinking about applying to the Lab?

The Lloyd’s ecosystem is complicated, so the more preparation the better. Companies should do research to become familiar with the various players and choose a good mentor. It’s also important to get access to information within Lloyd’s, which can be challenging. Once a company has decided who it will  be working with, it needs to get an NDA signed. Less is more, so try to limit scope – building a meaningful POC (Proof of Concept) in 9 weeks is a big task. Also, keep an eye on demo day as it comes by quickly. Last but not least, have lots of fun.

Dave Connors, Founder and CEO, distriBind

Can you give a definition of a bordereaux?

It is any record of the risks that have been written by a coverholder or MGA. It records the premium that has been paid or the claims that have been incurred that is passed from the agent selling the insurance policy to the insurance company providing the underwriting capacity, usually once a month 

What is distriBind doing?

We don’t think bordereaux should have to exist because they are very inefficient. We want to remove the need for them. We’re able to do that if insurers have digital distribution with their MGA or broker. It’s also possible if for claims they have a digital TPA (third-party administrator) or DCA (which is a TPA, as described by Lloyd’s), so we can receive data in real time. Where bordereaux can’t be avoided, we’re able to process various types and formats of data without additional manipulation. At distriBind we take a different approach to other bordereaux solutions on the market – we treat variances as an engineering problem, not a data problem for the user to have to sort out.

It’s in everyone’s interest to provide better data more quickly – why hasn’t the market been able to do this?

I think there is still a technology barrier. There has likely been the occasional one-to-one relationship where a coverholder is plugged straight into the insurer, but there hasn’t been any central facility for doing that. I think the biggest barrier has been the expectation that there needs to be a one size fits all approach. At distriBind we don’t require this. Digital coverholders are able to send insurers real time data, whilst more traditional ones might still rely on spreadsheets. This doesn’t have to be changed – distriBind offers a flexible way of handling data.

Does distriBind now have some real clients?

Portabl is our first client – the company will send data to its capacity through a combination of real time exchange and distriBind producing the bordereaux that the carriers need automatically. We’re also expecting to implement with a few other MGAs, including some more established MGAs. On the carrier side, we have a few conversations going at proof of concept level.

Is distriBind focusing primarily on claims reporting rather than premium reporting within the Lab?

For this initiative, we’re focused on claims because it is one of the Lab’s themes. But one of our mentors is a pricing actuary and another is an underwriter, so we’re looking at how we can present a unified portfolio view. Our risk and premium processing is ahead of our claims processing, so part of joining the Lab was to help accelerate our claims development.

Jeff To, CEO, Safekeep

What is subrogation?

Subrogation is the process by which third party liability is identified in insurance claims. For example, when a car accident happens and the claim has been paid, the process of subrogation begins when the first-party insurer determines who was at fault. If the other driver was liable, then that car owner’s insurer will pursue the other insurer, send out a demand and try to recoup losses that were paid out to the first-party driver.

How does Safekeep use technology and data for subrogation?

Subrogation is a time-consuming and labour intensive process – it requires a lot of expertise to detect which claims have recovery potential. Carriers are also dealing with a lot of unstructured data. Safekeep helps to deal with this data through natural language processing (NLP). We also apply predictive models to identify what claims have potential for subrogation –  and we help to automate the overall workflow, resulting in cost savings. Safekeep has been able to find a 20% uplift in net recovery potential ($) for our customers. We’re working with over 8 carriers, and Root has recently selected us as its strategic subrogation platform.

Do companies start working with you in a select line of business to test you out?

We learned the hard way that we shouldn’t deliver an all or nothing platform, because carriers have varying degrees of maturity in their subrogation processes. Safekeep works in auto, workers’ compensation, property, general liability and group benefits. The first step for a customer is to choose one of these lines of business. Step two is to scope it out, and decide how far back into claims data we’re going to look – usually it’s one or two years. The third step is to put us to the test. 

How are you working with the complicated data that’s coming through Lloyd’s?

There are a number of problems we face with the data that we get. Firstly, de-duplicating records has to be tackled. Another thing is extracting key information from notes and unstructured documents. Every client has varying degrees of maturity in its data – sometimes it’s really clean and we can get it through API, other times we have to do a lot of data ingestion.

Robert Hogarth, Co-founder, SettleIndex

What is SettleIndex doing?

We focus on disputes, particularly litigated disputes. We estimate that 95% of disputes eventually settle out of court, but they take far too long to reach this point. There’s potential for big savings to be made on legal fees. To do this, SettleIndex’s solution doesn’t cut lawyers out, but it harnesses their skills and experiences to help them be more efficient.

Can you explain how SettleIndex’s solution works?
We provide a framework to do expected value calculations in a way that captures the data. A rich source of data about disputes builds up quickly. Working in the Lab, we’ve found that a lot of settlement data doesn’t exist. One of the first things that needs to be done is to start capturing it – and our solution does that. 

We populate our model with data as our clients start to use it. We propose that insurers put all of their litigated claims through the system, and within a few years there would be a lot of data that currently doesn’t exist. This would give information such as the typical costs of claims and how long they take to settle. We also have some proprietary ratios, so we can compare one claim with another on a standard index to see whether it has been dealt with efficiently or not. We think that we’ll be able to get insurers to share their data anonymously.

How is the Lloyd’s Lab experience going?

We have five mentor syndicates: Beazley, Apollo, Liberty, Cincinnati and Chaucer. All of them have been really helpful. We’re close to doing a formal trial with a non-Lloyd’s insurer, and we’ve been through so many obstacles and people to get there. But with the Lloyd’s Lab, we got through all of that within two weeks.
 

Stay up to date with the latest from InsTech