Insurance Edge went along to meet LexisNexis UK & Ireland MD, Jeffrey Skelton and CEO
Bill Madison, to find out more about how data is truly driving most of the big changes
transforming the insurance industry right now.
IE: At the start of the 1980s hardly anyone had a mobile phone, but by the end of that
decade there were mobile networks, rival carriers and thousands of units sold. Insurance is highly automated now, data management has evolved rapidly but in some respects we are still in the early stages of transformation. Where will we be by the end of the 2020s?
BM: Our strategy is to help our customers in the insurance industry market their products and services more effectively to the end consumer. You’ve got data from first point of contact, quote details, assets being insured, occupation, lifestyle – all kinds of data sources.
I think in the next decade we are going to see a lot more of that, but on the other side of
the equation, there’s this thing we call the mill – in other words, the car, the vehicle. The
knowledge that’s coming out of the vehicle is going to increase and the way we get
information is going to change incredibly. I guess the next Chevron in that journey is the way that policyholders are also becoming consumers within the vehicle. It could be the safety features inside the vehicle, how often they’re being turned on or off and so on.
A great deal of the transformation is to do with how data insights and knowledge affect the efficiencies in process. So for example about 96% of the admin costs for insurance transactions are linked to processing data, maybe four or five percent is fraud. The way data is handled, the way insurers engage with consumers over their data, is just at the beginning right now. We are going to see a lot of change in the future.
JS: We have a crystal ball to an extent, in that the US market is slightly ahead of the UK
market. It’s the same as regards workflows, data plays a role in cutting expenses involved in that workflow, or improving the accuracy. In the UK we’ve launched a database so you can prove that you had car insurance previously, although that same database system has been around for 15 years in the USA. But all this sharing will enable a 360 degree view of the policyholder and there is an explosion of data ahead of us all.
The next step after quoting via shared or contributory data, is going to be claims. That can be used in so many ways, and everyone will get there because everyone has a
smartphone in their hand now and so we have democratised the whole process.
BM: The other side of the coin in all this is; what’s in it for the consumer? We see every
carrier being able to offer a reward to the consumer. Firstly, the process of getting
insurance will reduce to a second, from something like 45 minutes. We’ve already seen
that happen in the US Life market, so the benefits in time saving to the insurer – and the
consumer – are there.
IE: Once you know everything about a modern vehicle; where it was made, first registered and leased, geo locations, service records etc. and then you add that layer on top of the driver history, then you get to a point where you can personalise that quote much quicker.
BM: We are starting to see a data strategy develop with the insurance industry, and the
car manufacturers. We are going to play a role in bringing these worlds together. Now the question that arises from all this is consumer consent – how much do people wish to share online?
Now that’s something we are working on all the time, by building on a transaction, so we can see – insurer and manufacturer can also see – exactly when consent was given by the consumer and at what stage of the process. Everyone has to be compliant of course, in every marketplace, and those consent rules will vary in the future.
JS: It’s not just the OEM data of course. When you think about Open Banking and how
that has changed in the last few years, then you wonder `can we bring that same concept to insurance, for all the pieces of data that are available?’
IE; This brings us neatly to Alexa and voice search. If a consumer has a kind of digital
consent passport then that agreement covers a lot of moves on a consumer chessboard.
One shared consent interaction, verified of course, can save the consumer having to
repeat the same process with every company they meet online?
BM; Yes, they can have it in a managed state, by market. Those changes are definitely
coming and LexisNexis can play an incredible role in developing that in the next few years.
JS: The concept is pretty simple; `Hey Siri, give me an insurance quote,’ and that’s already developed in the USA. The backbone of that answer is gathering up all the underlying data that can affect that final quote. Right now in the UK, a bit of that data is still having to be added manually to get to the quote, it isn’t quite fully automated so we have to cross that barrier.
Another cool thing that is taking place is text-to-quote. Great, but you have to have all the
data, make sense of it, get it to the insurance company in a second, verify that it’s accurate of course – and then you can quote. We are not there yet, because there’s still too much data being asked of the consumer. But it’s coming. I saw a demo, where you could stand in a BMW dealership, look at a car, text BMW Finance and get a quote back within seconds. No more filling out forms in the dealership, no phone calls and Q&A, passwords etc.
BM; Lots of challenges surrounding insurance is the complexity of the product. So we ask
ourselves, can automating the data actually make everything clearer? That’s something
we really want to focus on more. If you look at various surveys, you see that some of the
least trusted companies are those within insurance and that’s partly because people feel
they’re not sure they’re buying the right product. That has to change.
IE; Let’s mention something topical in the UK; floods. How can AI be used alongside
historical data, to predict how localised flooding might happen in future?
JS; That is a tricky one because you’re tracking all kinds of trends and trying to model the
future. We have a platform that helps navigate those waters – pun intended. You can take government information, weather info, flood warnings, geography of a locale and put all that into a platform. I guess it is fair to say the data is there, but it hasn’t been fully explored yet. If you look at the most recent heavy rain event and then look at what happened afterwards; homes affected, losses, claims, repairs – everything.
IE; You have to build in all the planning applications too; new builds, change of use on
commercial, all kinds of stuff?
JS; The data is there, mostly, but let’s look at this from an insurer point of view: Premiums are historically low, you have people insuring a £250,000 asset for under £200. That same consumer might pay £400-£600 to insure a £6K car. There is so much more you could do to analyse flood data, sweep through all the planning, claims, housing type, local authority flood defences, river management etc. but could you do all that work and sell a policy at a price the market will bear?
The short answer is you could insure almost everything on the second floor and above a
great deal easier. Then there are cars being flooded in the same events. There is a great
opportunity to cross-pollinate all that house and business flooding data with vehicle
insurance because one or two cars are being flooded on driveways, so in theory you could text an alert saying `hey, maybe move your cars to higher ground, flood warning issued?’
In the end, you wonder if insurers and brokers can make all those layers of data, and all
that research, actually cost-effective, unless home and contents rates rise.
IE; Maybe the government and the insurance industry can sell the value of that detailed
flood data project to policyholders? Home insurance is amazing value when you think
about it, but if the climate is changing then perhaps everyone has to accept that prices
have to rise in the longer term?
JS; I was shocked how cheap my Contents insurance was in the UK. Which is great by the
way – don’t want to go back to the States and pay more!
IE; One last question on fraud. Will automated data enhance human instinct when it comes to detecting fraud, or replace it?
JS; Probably both are needed. Certainly right now we still need people to look at quotes,
policy changes, claims – everything. The type of data we have now is not so raw, so basic.
Technology has given us new tools to analyse all kinds of things like whether a particular
mobile or laptop has been previously used in a fraud case – digital footprints if you like.
BM; The thing Jeffrey just said about connected cars is a good analogy here. It comes down to columns of data, two worlds coming together. The asset being used to get an insurance quote is worth tracking and understanding, plus transactions and credit history. The more you have to look at, from a variety of sources, before you quote, the better. It’s gone way beyond checking someone’s postcode and the stats associated with that one piece of data.
IE; Interesting and educational guys – thanks for your time.