David King: Founder of Artificial Labs: Organising the world's data

Twitter icon
Facebook icon
LinkedIn icon
Podcast

After honing his skills and gaining experience in other industries, David King settled on insurance and, with co-founder Johnny Bridges set up Artificial Labs in 2013.

Today Artificial is helping deliver major efficiencies for insurance companies in how they collect and analyse third-party data.

David joined Matthew on Episode 70 of the InsTech London podcast and topics include:

  • Why the insurance industry is a great place to launch a business
  • The benefits of different accelerators and incubators
  • Lessons learned from Ping An and the Chinese market
  • Identifying the right technology mix for different clients
  • When to bring in a CEO
  • The importance of networking and relationship building

Summary of our discussion

You didn't start off in insurance. What attracted you to the industry? 

I'm a bit of a geek. I like data and technology and the applications of those. Insurance is an area where you can use technology at scale. There is the customer base, resources, and a competitive environment for firms. It has all of the components you need in order to use cutting edge technology because of the vast store of data.

You've got a bit of a history with InsTech London. Can you talk a little bit about what we did together?

When we were first trying to understand the insurance market, you and Robin were the pre-eminent people to try and talk to. To try and persuade you that we were worthwhile, we designed the logo and your first website, and that gave us some free or some sweat marketing in terms of exposure but also got us some insight into how the ecosystem worked.

What is the problem that Artificial is solving?

We're trying to help insurers digitally quote, bind and issue, to use all of the data that's available and provide their customers with a great experience.

Regardless of how they get sent the data, we help them quickly structure it in a digital format to enable them to augment this with third-party and internal data sets to score risks. We use a variety of rules-based approaches and machine learning, so by the time it hits an underwriter they've got a very detailed picture that helps them understand what the likelihood is of them writing this business, and is it worth their time?

You seem to be reaching back into the past, where people are still sending PDFs and emails. But you're also looking to the future by going directly to the original broker, or even the original client, to get access to their data through their systems.

You have to be sympathetic to where the market's at. There's no point in having a car that could win a Formula 1 race in 2025 if you're in 2020. Brokers are still going to send emails and they're not always going sign up to a single platform. But also, why would they sign up to 20-30 insurers' platforms?

That's the way the market operates, that's the way they're going to send the data, so they have to be able to consume it the way that your customer wants to send it.

Have you seen that trend change since 2015 when you started working with insurers? 

There is a trend in technology. Whether it's InsTech London persuading people or the internal endeavours of businesses, the adoption of technology is increasing as it becomes more and more credible.

Another factor is the hardening of the market and people realising they have to improve loss ratios. Once people appreciate that there's a market imperative to reduce their costs, the technology is a key driver of that.

How are you using other sources to complement or enhance the data that's coming in?

We will access whatever data is relevant for a specific line of business. We can access legacy data through file transfers and our preference is always via API.

An example of that is AXIS, where they're insuring professional footballers across Europe. We plugged into their binder systems and some other core systems, but we also plugged into Opta to give them real-time insights around the players and the clubs, where they're playing, etc, which was really valuable for them.

Is part of your role to advise clients on not only the data they can access but also the quality of the data for enhancing or complimenting their own submissions?

Yes. Our data team is run by a qualified actuary and somebody else in the data team has a PhD in astrophysics, so we match academic analysis and simulations with the insurance insight.

We can run simulations and models for the insurers to show them how something will perform over time, taking their data into account.

Are you able to help companies in how they store their data, so they can cross-reference different silos in their own organisations?

We give customers their own account on AWS, so everyone gets their own data repository. Every time they agree a risk, or bind a risk, or plug in the claims data, they can then go back and look at all the variables.

They can look at all the characteristics of that risk across all the data sources and how that impacts the performance of the risk or the book, and how that impacts the re-selection part of the process.

You took part in the Ping An accelerator. What made you want to take part?

What was really valuable for us was getting exposure to the Chinese market. In China, things move fast and the technology adoption curve is impressive. We wanted exposure to that, and we got an enormous amount of data out of it.

They gave us a lot of data so I got really good at the processes involved in training models and simulations and how this data should be used in a real-life context for insurers. Culturally, Ping An made us think at scale and do things very fast, which is a great thing to try and permeate into the culture of the rest of our firm.

You also did the Cambridge Judge Business School scale-up programme. What were the benefits of doing that?

We did the Cambridge Judge Business School and the PWC scale-up, both of which were very educational. Whereas Ping An was very commercially focused, the Cambridge Judge Business School brought us up-to-speed from an academic point of view of running the business. PwC gave us really valuable insight into how to approach legal and accounting and operating within large firms.

You've also been on the Capita Scaling Partners programme and Capita is now a client for you as well.

They invested in us and now hold a minority stake. What that gives us is great access to their understanding of enterprise sales. They've got good access to the London market. They have around 40 insurers as their clients, they also work with life and pensions businesses.

They understand really well how to sell into these firms and a lot of them are buying manual processes or outsourcing manual processes at the moment. We want to remove those processes, so it's a great partnership between us and Capita in that we are digitising a lot of these things.

There are a lot of people out there who claim to be able to get data from different sources. How do you distinguish yourself?

We don't have our own OCR (Optical Character Recognition) / NLP (Neural Linguistic Programming) technology. We go out and continually test the market for what is the best technology to use, so we can switch if one takes the lead and gives advantages. What we have are tens of thousands, if not hundreds of thousands of documents to do the transfer learning and training on in order to be able to understand how to extract the submission data.

Google might be the best one day; Amazon might be the best the next day. We will use whatever gives the best raw data, and then apply our insight to get exactly what the insurer needs from it.

You’ve also got some breaking news to report, in the form of a new partnership?

We’re extremely excited to announce a strategic partnership with Chaucer. We’ll be helping to build their next generation underwriting platform, which will hopefully make their bind and issue process completely digital.

They’ve got great resources, insights and technical underwriting abilities and we’re looking forward to leveraging all of our key technologies with them.  

What's the balance you look for in recruiting people with industry experience versus people from outside with other skill sets?

We need a great blend of phenomenally skilled people that are domain experts and have insurance insight. The exact blend is a bit of an art.

Johnny Bridges and I are the co-founders of the business, so we are very much start-up guys. We realised we needed somebody with the insurance background and the relationships in understanding insurance firms to do things at scale.

We partnered with Damian Arnold, who's now the CEO. He was previously the International Chief Operating Officer of Direct Line. He brought an element of insight to the business that was difficult to acquire as start-up people. He can operate at scale with insurers.

Looking ahead to 2020, what are your views on the trends you see?

There's a really exciting trend in using technology to increase speed. If underwriters can provide brokers with quotes digitally and quickly, even across emails, APIs and websites, we are getting to a point where we've got a decentralised marketplace.

If a broker can send an email to 20 underwriters and get a price back within three minutes from all of them, they are increasing the efficiency. Increasing the efficiency of that market means the customer, whoever they are, will get a better experience.

How do your products sit within the workflow for people?

We see ourselves as sitting on top of the policy management system. We digitise the structure and make everything in pre-bind more efficient. When the data goes into the policy management system of the insurer, that's where they own it. We're not going to compete with a policy management system, as we don't have the depth there.

How do you integrate into the work that your clients are doing?

All our functionality and components are separate and API driven. Insurers can pick all of our functionality or just one or two pieces and integrate them with their existing technology landscape or their planned technology landscape.

We realised very early on that every insurer is different and requires a different set of functionality. We make it as easy as possible to plug in with our technology, and our vision is that we would integrate with any platform in the market that is being widely adopted.

How do you manage to keep up to date with everything happening out there in the industry around you?

I listen to podcasts on the way to and from work, but my main thing is finding people that really understand the market and asking them hundreds of questions.

I was fortunate enough to meet Robin and yourself early on. And there are lots of other people in the market that are very intelligent and very well-informed.

We're thrilled to have Artificial as a core member of InsTech London. What is it that keeps you coming back to what we're doing?

The newsletter is great, but the fact that somebody like Google turns up to support at an event, and we can go there and listen to what they've got to say is also good. At the same time, you meet people from across the insurance industry, gather their insights and network. It’s a great community to be part of.

More information on Artificial Labs is available from https://artificial.io/


Continuing Professional Development

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.

Complete the InsTech London Podcast Feedback Survey to claim your CPD time.