Skip to main content
Article

Insurance technology in a post-COVID world: a watershed in the art of the possible?

Insurance Consulting and Technology
COVID 19 Coronavirus||Insurer Solutions

February 12, 2021

Willis Towers Watson’s first ever virtual expo for insurers, ‘Empowering growth and resilience’, concluded with some of our experts answering questions raised during the three days of the event about the use of technology in the industry.

In a year when the COVID-19 pandemic has changed so much about how work is done, it seems apt to have closed it out with a technology-enabled ‘virtual’ conference focusing on the role of technology in insurance.

Looking back over the past few months, the global response to the pandemic has demonstrated the expansion of the art of the possible – with organisations integrating and using technology to transform workplaces and operations more quickly and effectively than perhaps most thought achievable.

For insurers, many of whom were already on a pathway to a more digital future, COVID-19 has typically accelerated their progress. And now the train has gathered pace, the digital transformation of insurance that’s been widely discussed for many years, looks all the more imminent.

As the sessions from the three days of the virtual expo showed, there are many aspects to that transformation that raise many questions around everything from data strategy to disclosure. Our panel fielded a selection of questions related to key and recurring themes from the expo.


Our panel:

  • Duncan Anderson (Questionmaster), Global Insurance Technology Leader
  • Alice Underwood, Global Head, Insurance Consulting and Technology
  • Claire Farrelly, Global Leader, Multinational Client Excellence
  • Joel Fox, Global Product Leader, Life Insurance Technology
  • Charlie Kefford, Global Product Leader, P&C Insurance Technology
  • Katey Walker, P&C Sales and Practice Leader, The Americas
  • Evariste Yeung, Insurance Technology Sales Leader, Asia Pacific

Jump to questions:

  1. What for you was the biggest effect of COVID-19 on the way insurers work?
  2. And what lessons do you think we can take and apply to potential future economic shocks?
  3. The session about the role of Chief Information Officers talked about the need to build a digital mindset. How do you suggest you go about that?
  4. How significant do you see analytics in claims becoming in five years?
  5. Will artificial intelligence (AI) kill the underwriting profession?
  6. Should I start on the workflow or the data assets? More generally, what can I do today to prepare and adapt?
  7. Will automation make it harder and/or change how those starting out in their careers learn about the technical details of the business?
  8. Is automation really a worthwhile investment for a PC capital modelling team?
  9. What should we be doing right now to ensure we are ready for IFRS17 implementation?
  10. Across both P&C and life, which is the most common definition of fairness on life pricing that you’ve seen and/or are companies actually trying to address multiple definitions at the same time?
  11. There’s a strong imperative to reduce costs in our organisation, but there’s little bandwidth to undertake major projects and we’re already over-stretched so can’t reduce people or IT. How do we meet the challenge?
  12. We saw the presentation on your Strategic Asset Allocation (SAA) tool, but how often do insurers tend to re-optimise their asset allocations and what constraints do they typically work within?
  13. Finally, of all the technology trends discussed during our virtual expo, which has the most potential to change the insurance landscape?

Q1. What for you was the biggest effect of COVID-19 on the way insurers work?

Joel: I think the main thing was the need to assess the ‘new normal’ and what that meant for how work takes place and things like collaboration tools. It’s accelerated the trend around automation, and I think it’s fair to say we’ve seen a strong correlation between higher levels of automation capability and resilience throughout the pandemic.

Katey: Yes, I think you see that in the way some insurers have adapted to the changed circumstances and identified key performance factors that they entail. From personal experience, my mother collided with a deer in her car very recently and the claim was dealt with remotely in a couple of days.

Alice: That just goes to show that while insurance is not an industry known for changing quickly, it has shown it can. I think that modernisation efforts that have perhaps been bubbling away in the background will meet less resistance in the future.

Q2. And what lessons do you think we can take and apply to potential future economic shocks?

Charlie: Picking up on what Joel said already, I agree that companies that were able to quickly get to grips with what was going on, how it affected their business, and react in an agile manner have generally fared better and maintained better customer relationships. I think you’ll see that being reflected more in target operating models and the technology that supports them.

Alice: Yes, agility is huge in my book and chimes with one of my rules of thumb of risk management – that when looking at past experience to inform assumptions, know that there is something different coming. That’s why scenario testing and business continuity planning will continue to be very important.

Q3. The session about the role of Chief Information Officers talked about the need to build a digital mindset. How do you suggest you go about that?

Joel: For me, the key thing is to set the tone and direction from the top. The fact that many insurers have hired or appointed Chief Technology Officers is, if nothing else, a signal of intent. One important point I would make though is that technology shouldn’t drive the operating model – so the digital mindset is about building awareness of what is possible with technology and having the relevant skill sets to utilise that in the context of the chosen model and to implement the vision.

Claire: Yes, and allied to that is the need to think holistically about the technology ecosystem that will support the operating model now and in the future. What I mean by that is the connectivity, use of data and uses of technology so that insurers don’t have to keep thinking about the technical infrastructure but can instead focus on risk and their unique value add.

Back to top

Q4. How significant do you see analytics in claims becoming in five years?

Charlie: We see it increasing massively already in the P&C market, with variations according to the nature of the business. In personal lines, you’ve already got machine learning contributing to case estimates, fraud prevention and decision support on more complex cases. Those sorts of applications are also moving across more into commercial lines, which is also generating wider applications in areas like text recognition so that certain key words in a document are identified and surfaced to decision makers. Where we see future near-term potential is in more granular individual claims and policy level reserving and in enabling workflows that feed consistent claims data and insights into other parts of the business, such as pricing, underwriting and portfolio management.

Alice: The ability to interrogate very large datasets is helping to remove artificial function siloes in general. When supported with an integrated process, we can avoid, for example, different departments working with different assumptions.

Q5. Will artificial intelligence (AI) kill the underwriting profession?

Technology will, we believe, enhance the role of the underwriter, but put a greater emphasis on understanding and applying AI generated information, and combining that with their own knowledge and experience to reach better underwriting decisions.”

Evariste Yeung
Insurance Technology Sales Leader, Asia Pacific

Evariste: In short, no. Technology will, we believe, enhance the role of the underwriter, but put a greater emphasis on understanding and applying AI generated information, and combining that with their own knowledge and experience to reach better underwriting decisions. The reality is that there is going to be a place for complex human judgement in most facets of underwriting for the foreseeable future.

Katey: There’s a huge variety of underwriting needs, even within companies. I think there’s still this false notion that robots are going to take jobs, but I see AI as creating capacity and mindspace to support underwriters to do their jobs better.

Charlie: All I’d add to that is that people shouldn’t fear change. It just so happens that many of us on this panel are insurance actuaries. Twenty years ago, you didn’t see too many actuaries in insurance companies, but by embracing change and constantly improving what we do we have added more and more value to our companies. As a result, there is increased demand for our skills and far more actuaries in insurance than there were before.

Q6. Should I start on the workflow or the data assets? More generally, what can I do today to prepare and adapt?

Charlie: For me, data is critical. I’ve heard it referred to, cheesily, as the ‘new oil’. Get your data strategy right as a first step. Consider the role of automated workflows to support that strategy, such as enforcing a common data schema or capturing certain data automatically for a given task. Katey: Ditto to that, but what I’d add is that in the context of a transformation project that might have multiple phases, don’t let either data or workflow run too far ahead of the other.

Back to top

Q7. Will automation make it harder and/or change how those starting out in their careers learn about the technical details of the business?

Claire: A lot of automation in insurance will target the simpler, repetitive jobs that junior staff may have been given to learn their trade in the past, so we will probably change the way they learn the basics. But it should lead to more engaging roles – so instead of lots of cutting and pasting of data into spreadsheets that many of us may have had to endure, there will be opportunities to develop knowledge of areas such as data visualisation and automated processing.

Joel: It’s important that we don’t allow those starting out in their careers to lose sight of the fundamentals of what we do. But it’s a mistake to assume that automating a process makes it a ‘black box’. Whilst constant repetition of some basic tasks is how some of us may remember gaining a degree of our learning, the real learnings came from thinking through the underlying problems, and in the analysis, not from the repetition of the necessary steps in between. I can say from personal experience that manually changing parameters of hundreds of model runs isn’t what helped my technical knowledge and understanding of the results.

Q8. Is automation really a worthwhile investment for a PC capital modelling team, as systems and data formats and more are continually changing, so we will have to continually change the process?

Evariste: Maybe think about that question from the standpoint of how to get information into and out of models, given that your people are always going to be under pressure from change. Done well, automation will give you flexibility. By setting up the modelling process to evolve with market practice, you can bite off goals in manageable chunks while building in the flexibility that will service subsequent needs, say, for example, in the data capture process and methodology.

Katey: I would describe automation as becoming ‘table stakes’ for modelling. In our experience, the return on investment can be almost immediate, and set the foundations for future, iterative and meaningful developments.

Q9. My small organization has only recently begun thinking about our solution to the challenging new IFRS17 Standard. We were hoping that the large insurers would sort through the details and pave the course for us, but that hasn’t materialized. What should we be doing right now to catch up and ensure we are ready for implementation?

Claire: Self-serving as it may sound, I think insurers in that situation should engage an experienced advisor and consider IFRS17 solutions available in the market: “buy” rather than “build.” In some cases, you could have a full IFRS17 implementation in weeks.

Alice: Just because larger insurers haven’t shared their IFRS17 lessons to date doesn’t mean they’re not doing anything. They certainly are. The benefit of our and other solutions – and there are other products out there of course – is that they contain insights from working with a range of insurers.

Back to top

Q10. We heard about issues around fairness in the session on life pricing. Across both P&C and life, which is the most common definition of fairness that you’ve seen and/or are companies actually trying to address multiple definitions at the same time?

Duncan: This is a really tricky area. There are multiple definitions of fairness and you can prove mathematically that one of them will always fail – which can be manna from heaven for some journalists. One simple measure is “actuarial group fairness”, which basically means that premiums reflect the underlying risk. This can be relevant in markets that wish, for example, to limit price optimisation approaches, but it doesn’t limit risk segmentation in any way. Where the use of particular rating factors is not allowed, for example gender in the EU, the most common approach is “fairness through unawareness”. This means that a particular rating factor is either not gathered, or is not used in any analysis, which means the resulting rates don’t consider it at all. The problem is that most models will establish proxies for those omitted factors, with the result that when rates are summarised by the omitted factor at a simple high level, they still appear to differ, thus failing a further fairness measure of “demographic parity”. In practice, perhaps the most common concerns are addressed through “fairness through unawareness”.

Joel: The life market is less sophisticated, and most are working implicitly on the basis of fairness through unawareness. But the situation is no less tricky as, for example, in the case of gender, that could very easily creep back in ‘by proxy’ through other factors – quite likely where any medical information is being used, or occupation, or even benefit size which may correlate with gender for some life products.

Q11. There’s a strong imperative to reduce costs in our organisation, but there’s little bandwidth to undertake major projects and we’re already over-stretched so can’t reduce people or IT. How do we meet the challenge?

Automation is not just about cutting costs, it’s about creating bandwidth to spend time on more growth-related initiatives.”

Joel Fox
Global Product Lead, Life Insurance Technology

Joel: I get asked this a lot. Automation is not just about cutting costs, it’s about creating bandwidth to spend time on more growth-related initiatives. Don’t assume you always have to set up an enormous programme. We’ve worked with companies with big ambitions for automation but which have started small – creating the bandwidth for people to look at other things but also, crucially, generating buy-in for the changes. That last point is really important in my view, as success doesn’t tend to come if it’s done to you.

Evariste: A good starting point is business as usual, but you don’t necessarily want to only do the same thing cheaper - but also faster and better. That could be something like moving computing resource from in-house servers to a flexible, on demand cloud-based grid.

Katey: I think of it as “what is the cost of staying the same?” Maybe it’s a cost of missed opportunities or of not keeping up with the competition. You’ll always be too busy if you don’t make a start.

Q12. We saw the presentation on your Strategic Asset Allocation (SAA) tool, but how often do insurers tend to re-optimise their asset allocations and what constraints do they typically work within?

Joel: My experience is of companies doing deep dives every three years or so, although best practice would probably be every year and after major market events, such as we have now with the pandemic. Typical constraints might be some form of total risk measure constraint, volatility, duration constraints, or credit rating constraints.

Back to top

Q13. Finally, of all the technology trends discussed during our virtual expo, which has the most potential to change the insurance landscape?

Alice: For me, more and better data is more useful than developing a fancy new calculation. The processing power to deal with much larger data volumes, and the ability to integrate that data and generate insights from them, is where I see huge leaps originating from.

Charlie: Amplifying Alice’s points, I see that enhanced data capability really breaking down siloes within insurers. The biggest thing for me is the ability to pull out real-time insights and create value from them by delivering them to decision makers in all insurance functions.

Katey: I’d endorse both Alice and Charlie’s points for the reasons that everybody wants the answer now and, increasingly, organisations want, and need, a single version of the truth.

Joel: It maybe sounds like I’m avoiding the question, but I genuinely think it’s no one technology but a combination of all the technologies we have discussed over the past three days that will be totally transformative.

Claire: I’d agree, and think it comes back to the digital mindset we talked about. I’d caution insurers about putting something new and whizzy in a specific area in advance of stepping back and thinking about the whole technology ecosystem that will serve their future objectives.

Evariste: That’s consistent with the key point for me about any technology programme; agility. COVID-19 has shown us: expect the unexpected and understand that working practices that we thought were hard-wired can in fact be changed very quickly.

Links to recordings of the Expo presentations are available from your Willis Towers Watson consultant.

Contact Us