‘It’s part of our DNA’: How insurers are adapting to a data-centric world (and where they might be going wrong)

Automation, AI & Data Digital Innovation Panel Insurance

It is often said that data is the oil of the 21st Century. Yet, it is the digital engines – those artificial intelligence, machine learning, and API platforms – that will drive fundamental changes to business operating models and create new, enticing, and perhaps even controversial consumer services.

With the explosion of telematics, IoT and predictive technologies, many anticipate the insurance industry will be among those most radically transformed by these data-enriched (and data-enriching) tools. However, the ability to most practically and ethically utilise these platforms could prove the difference between those that thrive and those that sink in the rising tide of their data lakes.

Data and technology chiefs from Australia’s general and life insurance sectors beamed in remotely to our Future of Insurance 2021 conference to share their insights on the (often unrealised) potential of AI and ML technologies, the importance of API integrations in sustaining business-critical partnerships, and why progressive hyper-personalisation could potentially dismantle the very “historical foundations” of insurance.

We present highlights from the Digital Innovation Panel (Automation, AI & Data) featured at the Insurance event, with contributing insights from:

  • Simon Spencer, Senior Principal – Data & AI Strategy, Suncorp Group
  • Florence La Carbona, Head of Data, MetLife Australia
  • Amar Roomi, Chief Technology Officer, Blue Zebra

Moderated by Rita Yates, Chief Executive, Insurtech Australia.

Rita Yates (Moderator): We’re hearing frequently about how technologies like AI and predictive analytics are shaping the claims and the underwriting space. How are your organisations leveraging these innovative technologies and perhaps fast-tracking the end-to-end process you have in place?

Amar Roomi (Blue Zebra): I’ve never liked using the terms AI and ML as I feel they’re a little hyperbolic. But we’ve used these tools quite a bit in our organisation on the claims side and the typical stuff around triaging and cost sizing for claims from an underwriting pricing perspective.

We’re probably one of the few that have actually deployed machine learning algorithms directly into a live pricing environment. So it’s critical for us to continue to compete against the larger players, because our data capacity is still pretty small, though growing over time, and those machine learning techniques actually help us continually adapt our go-to-market, whether it’s pricing or claims servicing. And, really, when I talk about claims servicing and use it in that space, our triage process is huge because the end outcome for us is ‘How do we fix people’s problems quickly as possible?’ And part of that is directing customers to the right pathway. It’s a really big part of what we do.

Again, we don’t necessarily think of [AI] as a sexy tool anymore; it’s just part of the arsenal that’s at our disposal.


We’re lucky that our underlying data doesn’t have any limits or constraints; that’s quite useful from our perspective. But it’s a big part of our DNA.

Florence La Carbona (MetLife Australia): I also believe that it’s definitely going to be part of our day-to-day. We’re still the pioneer in a sense of paving the way of using these tools and capabilities.

To me, it’s more about finding the right use case and the right problem to solve to be able to demonstrate the value of using this tool.

By doing that, the end goal should be to demonstrate the value of solving a problem, rather than having the end goal to put something in production at all costs. And what are the learnings that we can get out of experimentation?

Last year at MetLife, we worked with our global teams, our claims teams, and technology teams to work on return-to-work and return-to-health models for claims. We’ve learnt an enormous amount. We know that this model might not go into production right away, but the work that we’ve done can be used in the future once we have the right infrastructure. We have identified the type of data that we need to capture to be able to run this model in a more effective way and to get this model to learn better and to be more accurate.

It’s more about accepting that success doesn’t mean putting something in production, but about [understanding] how we can apply these learnings in the future.


Rita Yates (Mod): And just slightly extending that question around ensuring the ethical use of AI, which is more than just in insurance.

Simon Spencer (Suncorp): Florence said some really relevant things. I completely agree that you need to be able to find the right use cases. But as you do that exploration, it’s also about getting some of the foundations right as well. And some of those are technical foundations and around getting the right data.

Many organisations have great architectures that are well aligned with business intelligence but perhaps less well aligned towards supporting AI and machine learning.

And, in some cases, they’re well supported in generating great insights, but not always in operationalising and deploying those insights into actions. Some of the technical footwork has to be done, but many organisations also need to look at what they should do and what they shouldn’t do? How do we navigate that ethical landscape of delivering value?

We don’t want to be the casino that is trying to get the person to play the poker machine endlessly. We want to be the organisation that is trying to find ways to deliver compelling value to your customers.


This has been a debate that Suncorp and others have also participated in, which is looking at putting together the ethical practices for using data sciences and also, by extension, in artificial intelligence: Where should we operate and how do we also ensure that we’re always operating in a way that’s beyond reproach?


Rita Yates (Mod): Moving on to usage-based insurance, which really isn’t a thing of the past anymore – plenty of insurers are offering more flexible products by leveraging telematics and wearables. So, again, how are your organisations using these technologies to provide more personalised and flexible products?

Amar Roomi (Blue Zebra): For me, insurance plays two really pivotal roles in society. One is around risk management, which is essentially risk sharing, so that individuals and organisations aren’t faced with financial shortcomings come claim time. The other really important service that we provide society is risk mitigation. When I look at technologies like telematics as an example, I see an opportunity to help better educate the public at large about whether it’s driving behaviour, whether it’s mitigating risks around the home, be it reducing the event of a fire or reducing the potential for crime. That’s the bit that I’m actually passionate for us to do more of as an industry. And we already do so much, but it’s not necessarily well pronounced.

Building safety standards have a long history of being associated with insurance companies, driving the path forward and making sure that they’re a standard, as opposed to something that people just volunteer to adhere to.


So, there is a really important component between the use of telematics and internet of things (IoT) technologies and mitigating the risk faced by our customers.

Simon Spencer (Suncorp): Just adding to Amar’s points. We’ve done a number of things looking at telematics in terms of pricing, utilisation and usage and so forth. I’d probably characterise a lot of the work at the moment as exploratory in nature. But what we’ve learned from it is, again, it needs to be about value and, moreover, value to the customer. Much of that is around providing customers with flexibility and that business agility to be able to craft products based upon contexts.

A lot of it is about creating systems and technologies that provide the ability to ingest and use and utilise many different types of data: it can be geospatial data, imaging data, sensor data, drone data, all sorts of other data. And we have to bring that into a landscape so that the business can consume that data at a level of quality that’s sufficient to achieve good business outcomes that deliver value to customers.


And there are lots of challenges tied up in all of that. But the good thing is that we’re well underway in that journey.


Rita Yates (Mod): And, Simon, creating a digital platform and using APIs to integrate with partners obviously plays a key role in any modern transformation. How is that journey going for Suncorp?

Simon Spencer (Suncorp): The term digital transformation was very much in vogue three years back – and not to suggest for a moment that the job is done; it’s certainly not. But it was the buzzword roughly three years ago and it continues to be a very important theme now. But now people have pulled that term apart a little bit.

The industry has realised that digital transformation is more than just the front end.


It requires a number of things to be done and done very well. And the same goes for artificial intelligence and data sciences as well. It’s not actually enough; you actually need to tie a few things together very well. And so the digital transformation story on the front-end, the integration story, the API story, tells us we need to plug things together efficiently. Obviously, the data services story is about generating high-quality, reliable data that can be consumed by your data science and digital platforms. And then, of course, there’s still the emerging automation story, and whether it’s your robotic process automation or your more sophisticated AI automation or business rules capabilities, all of these things are facets of the same basic problem space.

Many organisations that have tackled [the implementation of automation tools] one by one have struggled a little bit. But those that are starting to fit these things together into a holistic story are actually starting to pull ahead.


And again, Suncorp’s got some great things to talk about there, but it’s early days in some cases and it’ll be a continuing journey.

Amar Roomi (Blue Zebra): I’ll touch on Simon’s comments around API integrations and the key role they’re going to play moving forward. I want to say two things, one is slightly controversial. Don’t let your internal technology teams stop you from exploring what can be done via integration. At the end of the day it’s data in and data out, and you’re going to make a decision with that data somewhere in the middle.

APIs, to demystify them a little bit, are actually really easy.


Jump onto YouTube, look at a couple of clips, if you’re so inclined; it’s actually not the hardest thing in the world to go and test out yourself.

And one of the things that I’d love to see more of in the insurance industry is a bit of democratisation of technology expertise. It shouldn’t be the domain of your IT world. Go and try it yourself; have a play and see what you can do in your spare time.

API integrations are going to be fundamental to the success of, I’m convinced, every insurance company going forward, because you’re not going to be the best at everything, you’re going to need to consume services from outside of your own organisation – and that’s the key platform to do it through.


Rita Yates (Mod): This moves more into data privacy concerns. Obviously, this will surface more in the digital age, and the question around how insurers or related entities can continue to infuse transparency and trust in their interactions with customers.

Keeping in mind that we know that trust is not or has not always been a word synonymous with insurance – and, in fact, rates quite low in the entire insurance industry – how can we actually make that better at a time when trust under more threat than ever?

Amar Roomi (Blue Zebra): It’s a great question and I don’t have the answer. This is literally the question of our upcoming century, which is around data privacy and what organisations can and can’t do with that data.

The reality is people are doing things with your data that most people are probably going to be uncomfortable with.


Looking at some of the things that are happening with bigtechs, especially in the US, a lot of this probably feels out of the reach of most individuals, because they don’t know how to put a lid back on the bottle, to a degree. The cat’s out of the bag as far as data and the privacy that you might enjoy. And even obviously sensible attempts to secure the population with Covid; if you look at it, without the virus component, it would likely be considered overreach by lots of different organisations, be it government or private, to know more about you and where you go.

So, I don’t have a great answer for it, in terms of how do we engender trust with our customers, other than to be pretty honest with ourselves to say that these are really difficult questions.

We’ve got to be constantly monitoring what we do with our information, treat the customer’s information as it’s our own, and it’s got to be a continual discussion at the C-level.

It can’t be something that’s just hand-balled over to someone else to deal with. It’s a joint responsibility right across all our organisations.

Simon Spencer (Suncorp): This obviously takes us back to the ethics question earlier, which is something I feel passionate about. Organisations have a choice they can make: they can decide that they will treat the customer as the product, which is obviously the Facebook or Google model, where you are the product essentially; or you can treat the customer as a customer. And, moreover, to continue to ask ‘How do I be relevant in a customer’s life, and how do I deliver value to that customer?’ Organisations that pursue a high ground strategy, that can build customer loyalty and customer trust and not disenfranchise people from their own data – and that’s certainly where I want to be, and I think many people also want to be – where you’re looking at the question of ‘How do I add relevance and add value to our customers’ lives?’

It’s an evolving story, but it’s also the best defence against bigtech and some of those disruptive models – that is, the strength of our relationship with the customer.


And if you go back to Florence’s comment earlier around customer-centricity – ‘What’s the true north of the company?’ – and that has to be about delivering value to customers.

Florence La Carbona (MetLife Australia): And we also have a responsibility to educate our people internally about these concepts, as well as our partners and our customers. For example, at MetLife, we really work closely with our partners to equip them to understand the value of insurance. We need, for instance, their member data so we can give them insights on the claims experience so they can deliver better service to their customers. And being transparent about what is it that we’re going to do with your data and what’s in it for you – whether it’s for our partners or our customers – is really important.

And the same goes with pricing. Of course, for pricing, we’re going to use lots of data about our customers, about their behaviours, their claims histories and so on.

But, ultimately, we really need to position that and educate and to make it simple for people outside the insurance industry to explain what it is we do with data and what we need to do to provide a better price, better product, and better service.


Rita Yates (Mod): Florence, with all of this extra access we have to data, and in the midst of a battle for data-driven analytics and insights, do you have any keys to success around hyper-personalisation?

Florence La Carbona (MetLife Australia): I’m not sure I have the keys to success, but I might have a few opinions or things we’re trying to pursue. One thing that was said before, it’s not all about the technology, it’s really about the people. And Covid was a great experience in that sense that, ‘Yes, we all have access to technology to talk to people, but – and I believe that we are social animals – nothing will replace the relationship and understanding people drivers, whether it’s customers or partners. And this is really the combination of product people, for example, really knowing their products and their customers and their partners and having intuitions on what should come next, working in collaboration with people who know the tech, who can manipulate the data to reinforce these intuitions or tweak this intuition.

I believe it’s mostly about strong collaboration between people, having different backgrounds and making sure that people who know nothing about technology start to learn about technology and technical people really understanding the business, so we can find common ground. From there, personalisation can really start.

Hyper-personalisation will mean different things for different organisations. That’s why I believe the value of the organisation is paramount, because it should really stick with the story that we want to tell our customers. And, at MetLife, we have our own philosophy, our own culture, and this is what should transpire in the personalisation journey.

Amar Roomi (Blue Zebra): Hyper-personalisation, if you look at it from a pricing perspective, is going to come up against the whole concept of insurance, which is risk sharing. At what point do you personalise a price so that some other people effectively become uninsurable, potentially negating the historical foundation of insurance, which was ‘We’re sharing bad luck across a community effectively’? And yes, there’s a moral hazard element there, because if you’re sharing the bad luck, then maybe the individual that’s a little bit more risk-taking is going to be diluting value across that community. But it’s a really interesting question, because at some point those two fundamentals are going to come crashing head-on, and we’re not quite there yet.

But I don’t know that we’re that far away from seeing a situation where pricing is so customised at the individual level that the concept of risk sharing becomes a little bit different to what we used to.


Simon Spencer (Suncorp): I do think that success in personalisation strategy does come down to getting that intersection right between the people and the business and the technology story, and then the overall strategy objectives or what you’re trying to actually achieve. And getting it right, you’ve got to have the right technical plumbing, but you’ve also got to have the right business strategy or, indeed, what you’re actually trying to achieve here.


Rita Yates (Mod): My next question is around insurtech, something very close to my heart. How can industry players support the growth of the Australian insurance technology ecosystem?

Amar Roomi (Blue Zebra): Yeah, that’s a great question, Rita, and obviously close to my own heart given where we are today. For me, the biggest secret to unlocking insurtech – and I don’t think of it as just necessarily external companies, but insurance talent within the Australian market – is to allow people from across the organisation to up-skill, whether it’s a technologist upskilling on the business or the reverse, the business upskilling on the technology.

We’ve got to democratise that skillset far more than we do today and create a lot more opportunity at entry-level and beyond to effectively empower our people to solve the very many challenges that we have, rather than outsourcing them to ‘experts’.

And my genuine preference is usually to experiment with this internally and see what you can solve yourself. And then, sure, if you think it’s well outside of your area of expertise, reach out and get help.

But that’s one way to cultivate insurance talent in this industry. That doesn’t mean you’ve got to do it in-house. Please reach out to your colleagues within the industry as well, including within the insurtechs. But don’t think problems are insurmountable just because no one solved them just yet.

Florence La Carbona (MetLife Australia): And it also comes back to simplification. With digital, there are so many opportunities for simplifying, because the insurance ecosystem and an insurance business can really recentre on what we know best, which is insurance and insurance products, actuaries et cetera, versus developing software because we think that what we’re doing is really too complicated.

Now there are so many amazing solutions that we can leverage out of the box, and we don’t have to be software developers at the same time as being experts in the insurance business. So I believe there are huge opportunities for partnerships, and that’s really the way to go in the future. And, also as an industry, we all learn from each other as well.

Simon Spencer (Suncorp): I made a comment earlier that summarises it, which is:

You don’t have to be great at everything. Figure out what you’re good at, and then find people who are really good at things that you’re less good at.

In that lies the opportunities for large organisations to collaborate really well with insurance techs and other partners who have capabilities. The API story is also a powerful enabler for collaboration between large enterprises and smaller firms. And certainly, Suncorp’s doing a bunch of that. ◼

This is an edited extract from the Digital Innovation Panel featured on Day Two of FST’s Future of Insurance, Sydney 2021 event.