How technology is positively impacting the insurance process
BASS Underwriters’ Joe Good – Actuarial Analyst, and Hiscox London Market’s Paul Carnegie – Line Underwriter, Commercial Property, share their thoughts on the impact technology is having on the insurance process for insurers and their coverholders; from the pricing of risk, through to control, and the value of real time data.
Mixing the traditional with the digital…
Joe Good: “No one is walking into a building in Wall Street and trading with paper slips anymore, why aren’t we doing the same thing with insurance policies? Why is everything digitised in the finance space but if I want to bind a policy in Lloyd’s I have to gather all these papers, stand in a queue, just to get it signed off? There have been some good steps forward taken, particularly as a result of covid lockdowns, but the whole process still has loads of room to improve – the technology hasn’t been applied to the insurance space like it has in other industries.”
“But, now, people are starting to flirt with it – we see companies like Lemonade – where it’s all technology-based underwriting. But you can’t just rely on digital distribution to replace the underwriting process. It can’t replace all the tradition and non-standard type processes that the industry has created over hundreds of years.”
“My perspective is you can use all these new tools – the digital distribution, the online platforms, the third-party data – as long as you have a team that also has the underwriting experience and knowhow. There are instances where you can eliminate the underwriting process and let producers bind away and use the actuarial science to steer the book – but when it comes to specialised or grey areas of risk, insurtech is not replacing that yet.”
“If you can couple both – tech and underwriting experience and knowhow – it’s a fantastic combination.”
Paul Carnegie: “Across the City of London there are algorithms and high frequency trading taking thousands of data points and making big decisions. A typical insurance underwriter on a big risk might ask 10-20 questions – don’t tell me that can’t be helped by technology? But you do need someone with strong underwriting expertise to control the machine.”
JG: “We use artificial intelligence to price risk that would take a whole actuarial team to do – it eliminates the needs for a team of five taking months to crunch out data for predictive pricing models. You can build these models in a short amount of time – two weeks or less. Cutting down time and resources.”
Keeping control of the underwriting process…
JG: “The control aspect is as important as having the technical rating. We have a lot of additional quality controls that couple the technology. One is just good old fashioned internal audit. I scan through inspections, pick out a couple and get a feel for what is binding and how the quality control process is going.”
PC: “One of the wins for Hiscox BindPlus – an online platform for homeowners and commercial property risks – is we can be as complex as we like because with an API it just tells the coverholder whether the risk is acceptable and what the price is. It gives the end coverholder confidence because they know it’s been through a set of complex rules.”
The value of good, real-time data and advanced analytics…
PC: “We have a large number of different delegated authority contracts. For traditional binder business, all that data comes through in a slightly different format. By the time the coverholder submits the data and we cleanse and model it all as one portfolio we’re three to four months behind. That makes all the difference especially in a transitioning market. The next stage of BindPlus is to take the real time data from BindPlus into Hiscox’s exposure management system. This will give us a real time view of our exposure and in force metrics.”
“But there is also a danger to real time data; you have to guard against knee jerk reactions. If you get data every week, how do you know if what you’re seeing is simply a short-term blip or not? It can be easy to over react.”
“The putting together of historical and ongoing claims and risk level data is also key. Capturing that data in the right format going forward is so important – even if you don’t know what you might do with that data yet. As an example, if we have a multi-location policy that has had a claim that a TPA has handled, there is nowhere in a structured data format to be able to tell which location the adjustor went to. It’s simply not been captured in a way where it can be mined. We are putting pressure on TPAs to produce the kind of data that we take for granted on the risk side. Risk level data has been in a good place for a while but the claims data is some way behind.”
“With the advancement and adoption of sophisticated data collection and better tools, we can look at portfolios in a better way, with more two-way feedback between insurer and coverholder, giving us the ability to make course corrections earlier.”
