#Insurtech is a launch pad to unimagined possibilities
Insurtech is where the insurance industry’s future begins, a starting point that is unfathomable, challenging and yet well within reach.
Think it, and it can be
A lot has been written about the state and use of big data and predictive analytics, the Internet of Things (IoT) and artificial intelligence (AI) in the insurance industry. Market commentators have been busy promoting the impact of technological advances across the insurance value chain for some time now and analysing how the sector is already facing unprecedented disruptive change. For all the noise, industry adoption and use of these concepts and technologies vary significantly among geographic regions, firm tiers, and types of products and lines of business. Disparate definitions also make it difficult to independently and comprehensively evaluate progress.
The reality is that while many large and multinational underwriters and reinsurers started their big data and predictive analytics journey over a decade ago, most midsize and smaller insurers have only recently started to transition from data management to true big data initiatives.
While there may be some debate about just how deeply advanced technologies and methods have penetrated the sector, what is unequivocal is that insurers as a whole are increasingly investing in technology, in analytics capability and in data. Change is anticipated, even if it can’t be described as wholesale at this point in time. Figure 1 captures the current reliance on data enrichment and acquisition, IoT data utilisation and AI deployment.
Figure 1. Sources of data generation explored by insurers:
Transformative tools and technologies, regardless of their adoption to date, will inevitably evolve and improve, and open the way for start-ups with a focus on #insurtech, such as Lemonade and Trov, to gain a foothold in the market. Equally, some insurers, notably Progressive and Vitality, are embracing change, partly through product developments that rely on telematics and wearable device technologies. In the process, these companies are developing increasingly loyal customers who value the broader risk mitigation services offered, while also gathering significant volumes of insurance-relevant data that can be used to better understand customer behavior, risk and need; they are well placed for future profitable growth.
In light of the threat posed by start-ups and early movers, how should established insurers react to keep pace? How do they evaluate the future technological, data analytics and competitive landscapes and proactively plan for them?
This will inevitably involve some stabs in the dark, but there are also some elements that provide a more predictable basis for action.
Certainly, the collection by insurers of unique data sets will be key in a more connected world. Smart, value-adding apps and the design of products and services that, by their very nature, generate lots of useful insurance and customer behavioral data, will underpin insurers’ attempts to gather unique data sets. This can help them learn more about customer needs, allow them to get closer to their customers and facilitate a new type of relationship that will shift from risk mitigation to risk management. However, it will also have very broad implications for the likely volumes of data to be gathered and processed.
Some companies have already started to store big data — both structured and unstructured — in data lakes, instead of simply organising them; they are then using technology to access and interrogate their data. Results have, in some cases, been impressive.
Moreover, we can safely assume that insurers will be able to extract more value from data. Already, AIs have the potential to extract information and generate insights in a similar manner to humans through voice and facial recognition, such as an understanding of stress levels in voice tone or through micro-facial expressions, decoding human emotions and using natural language processing. Their processing speed and computing power also enable them to analyse and recognise patterns and combinations that humans cannot even fathom.
As a result, AIs can find patterns and anomalies more quickly and efficiently, deal with unstructured data and data streaming in real time, and, therefore, improve the quality and efficiency of decision making substantially. Opportunities for insurers to combine big data with AI include cybersecurity, claim automation and fraud detection.
AIs also make real-time risk assessment and mitigation possible, and can detect, identify and prevent security threats by combining machine learning, text mining and ontology (relationships among properties of beings or things) modeling to curb illegal trading or fraudulent insurance claims as they happen. Some of the planned uses of technologies are referenced in Figure 2.
Figure 2. Technologies that insurers expect to have the biggest impact on the sector over the next two years and the next five years:
But, for all the excitement around these opportunities, we can anticipate there being some more fundamental questions for insurers to grapple with. For example, will micro-segmentation make insurance unaffordable for those who actually need it? And how will current risk mutualisation models that create the most efficient ways to share risk, evolve to account for more personalised products?
…a world in which your digital footprint gets recorded in real time into an immutable, highly encrypted distributed digital ledger (e.g., a blockchain) that is connected to your watch, phone, car, smart house, private virtual assistant and wearable health tracking devices that track your sugar and cholesterol levels, brain activity and molecular decay.
These data can provide insights into your driving and sleeping behavior, moods and stress, recently acquired taste for expensive wines, the extent of your social network or the fact that your running shoes have lost their grip. You own that data and have the private encryption key that allows you or others to decrypt it. Your private virtual assistant — an AI that has learned to defend your interests — communicates and negotiates with other AIs on how to use that data to, for example, get access to better and more personalised insurance products. Once granted access, the underwriters’ AI accesses unstructured data lakes containing the immutable record of your past behavior.
Based on the information, AI-enabled blockchain smart contracts self-issue a personalised insurance policy, designed only for you, with clauses, prices and coverage that self-adapt in real time to your needs by learning from your past behavior and assessing your riskiness in real time. This AI may be able to infer that despite the fact that you consistently drive at a speed that is 10% higher than speed limits, what actually constitutes a higher accident risk factor is that you have slept less than four hours or had a discussion with your preteenager during breakfast. Indeed, the moment you stepped into your connected car, your private AI assistant already let your car know that those two things happened and may affect your driving.
Your car may suggest to drive itself that morning or, if you refuse, since you are in a hurry and in a bad mood and think you’ll be able to get to your meeting more quickly than your driverless car, it will start communicating with the AIs of neighboring cars to make sure everyone gets to work safely. This will all be followed closely by your smart insurance policy, which may decide to charge you more or may decide to help you reduce your stress and incentivise you to change your habits. The wearable device tracking your brain activity will have the intelligence to let you know many hours in advance that you’ll be having a migraine or to give you tailored games to play on your smartphone to avoid certain parts of your brain losing its memory or certain types of synapses disappearing. It will let your smart disability insurance policy know about the changing odds of developing Alzheimer’s disease. AIs will combine information about your sleep, your geolocation, weather, trail conditions and your sneakers’ soles to assess the likelihood of a fall and communicate to your accident insurance policy. Even more, if we think about full-stack insurance, AIs could not only assess riskiness of individuals, but also automatically pool them to mutualise against those risks in large, decentralised, autonomous peer-to-peer insurance networks and find investors whose risk appetite is adequate to insure/reinsure them.
The need for scenario planning
Almost infinite data-driven and technological possibilities exist beyond these more predictable areas of #insurtech growth.
It is virtually impossible to guess what the future might hold for the sector, but that shouldn’t stop insurers from trying. Simply running through a sufficiently structured and informed future-state scenario planning process can help reveal new opportunities, strategic weaknesses and hedging opportunities that would otherwise be missed. It is important to consider a range of possible future states, from those that envisage little change (perhaps one that sees incumbent insurers shifting their operating models to accommodate a more digitised and data rich environment) to a much more extreme (arguably dystopian) view of the future. This breadth of thinking can allow insurers to develop strategies and build business models that are more adaptable and resilient.
Aside from what can be imagined for technology and analytics, there are, of course, a whole host of factors that will determine how the future will actually unfold that need to be built into thinking. Regulators will keenly watch the use of personal data and the role of insurance in a world that increasingly facilitates risk assessment that considers a “segment of one.” Consumers will be particularly concerned about cyber risks, which will be exponentially larger in a fully digitally enabled world. Furthermore, insurers’ access to the levels and types of data suggested in our Imagine scenario would appall many consumers, who might not engage with any insurer that hinted at such a prospect.
Nonetheless, every piece of technology needed for a world like that described in our Imagine scenario already exists. And some of them are likely to appear in some consumer segments and in some markets.
Regardless of the speed at which #insurtech and data analytics innovation occurs, the future of insurance is very likely to be one in which the nature of risk, the value of data and the relationship between underwriters and customers is radically changed. The impact will be felt on products and distribution channels, customer and risk models, regulatory frameworks, and the transition into cognitive big data and next-generation predictive analytics.
Like any other transformation, this requires strategy and direction, crucially underpinned by company views of the future state — despite the many uncertainties and out-and-out unknowns. Just as important is where the company fits into those future states, the associated implications for data and analytics, risk appetite, talent retention and attraction, customer relationships, market positioning, cost-efficiency, and the required levels of innovation and change.
Transformation won’t happen overnight, but has to start somewhere and incorporate some longer-term scenario and contingency planning to set a general direction. Subsequent practical steps to kick-start programs include:
- Look at fundamentals that include a foundation-level data infrastructure that is or will be in place and a corporate culture that understands the value of data and enables innovation.
- Implement a joined-up approach to opportunities and initiatives. Pick some that have medium-term revenue targets and would be ruthless in execution.
- Do something sexy and speculative. Lead by doing, as it will give insight, aid recruitment and retention, and may be the source of some lasting competitive advantage.
Opinions abound on the speed and extent of disruption, but there is general agreement that companies best able to ride the disruptive wave of #insurtech are highly agile, geared to execute managed risk-taking activities/investments, quick to make major decisions, and are capable of forming strategic/exclusive partnerships when advantageous.
A well-articulated vision of possible scenarios, backed by sound data and analytics foundations, and a streamlined strategy for bold execution will foster such qualities.
Magdalena Ramada specialises in emerging-market research at Willis Towers Watson, Miami
Andrew Harley is global leader of the advanced analytics practice at Willis Towers Watson, London
For any further enquiries please email Software.Solutions@willistowerswatson.com