Two Become One: The Convergence of Insurtech and the Industry
By Jonathan Spry, Co-founder and CEO, Envelop Risk
In the first part of a two-part examination into the cross-over between traditional insurance and tech innovation, Jonathan Spry, CEO and co-founder of Envelop Risk explores the data-driven future for the industry, the advantages of artificial intelligence (AI) and the cyber ‘shop window’.
Like many in the insurtech community, I am convinced that AI is the future of the insurance industry. Alongside traditional approaches, the application of AI could increase profitability (as underwriting and operational functions are partially automated) and provide insurance enterprises with a new toolkit which combines the best of human and machine expertise when making risk and pricing decisions.
The advantages of AI
AI and, in particular machine-learning, can be used to analyse large complex data sets; revealing predictive factors which are not easily uncovered by human analysis and conventional computation alone.
If used correctly, AI can overcome the human biases that impede rational decision-making and provide huge benefits over the purely human judgement model of uncertainty and appraisal of risk.
In practice, AI combines several disciplines including statistics, regression analysis, computer science, and human domain expertise which can replace or supplement nearly all previous forms of analytics. If these techniques were applied to insurance in a similar way as other data-rich industries, then AI could be a major force in driving transformation across the industry.
Man meets machine
The combination of AI with human intelligence is sometimes referred to as augmented intelligence. The application of augmented intelligence to underwriting (‘augmented underwriting’) is critical to meeting business objectives. However, the underwriter’s skillset will have to evolve to make the best of this resource.
Machine and human approaches complement one another. Human expertise informs model design and output interpretation while quantitative methods identify trends, predictive indicators, and dynamics that are either new or too subtle for human perception. The result is underwriting that takes the best elements of statistics, expertise, and common sense.
Leading insurtechs see themselves as architects and enablers of the future of insurance, combining the best of what is working now with the grounded, intelligent application of the right technologies. Augmented AI brings additional analytical discipline to underwriting (for instance, by not charging less than a modelled rate plus a margin for new business) but allows the continued benefit of using a specialist labour resource (in peer reviewing the modelled outputs).
AI is not a replacement for human resource, but, when used properly, is a function which enhances human decision-making and brings out the best from underwriters. In a new model of augmented underwriting, expertise will be needed to monitor, ensure quality, and share such insights with brokers, clients and management. It is becoming clear that none of this involves technology alone, but equally that an underwriter who cannot embrace AI is going to be left at a clear disadvantage.
Cyber: The testing ground
Cyber has become both a shop window into insurtech and perhaps the most obvious class of business dominated by a data-driven approach, on both the primary and reinsurance side. We have observed that as software eats the world, cyber takes a large bite of insurance, perhaps overtaking property risk in prominence over the medium term.
That said, the market for cyber-insurance remains imbalanced, with demand outstripping supply. The insurance solutions available contain strict monetary limits, well below potential economic exposure, as well as exceptions for certain types of loss.
The primary barrier to industry growth has been accurate risk assessments to expand coverage at appropriate pricing, further impeded historically by a yet fully formed reinsurance proposition.
The evolution of cyber-analytics technology and its role in the insurance eco-system is key to market development. Currently no dominant design exists for cyber-analytics or for cyber-insurance policy scope.
Anomalies are present in the market due to a lack of transparency and purpose-built products to match emerging demand. System-building and clarity is required for the cyber insurance market to move from an innovation/tech-transfer stage to full commercialisation, with a sophisticated cyber reinsurance product of critical importance.
Power of partnerships
To understand and price the nature of cyber threats, insurers need to access vast computing power and embrace techniques such as machine learning. However, insurers are unlikely to have such technology in-house and could look to gain access to these complimentary assets through partnerships.
If insurers are to truly embrace these disruptive innovations, then they will need to manage the risks attendant with collaboration, innovation and value-chain integration.
In the second part of this exploration into convergence, we will look further into the power of partnerships.
In the second part of his examination into the crossover between traditional insurance and tech innovation, Jonathan Spry, CEO and co-founder of Envelop Risk, explores how insurers and innovators can work in harmony…
To date, start-ups have been the catalyst for innovation in the industry. Funded primarily by venture capital, entrepreneurs have sought to offer novel ways to underwrite and deliver insurance products.
Digital disruption and division
Insurtechs have been on a mission not only to improve existing models but also to outcompete insurers with new and inventive products. As the visionary engineer and architect Buckminster Fuller said: “You never change things by fighting the existing reality. To change something, build a new model that makes the existing model obsolete.”
Many insurtechs, particularly in the US, are, in fact, seeking to disrupt the traditional insurance ecosystem rather than partner with it. While this model of disruption has proved successful in other industries, the insurance industry has thrown up high barriers to entry via stringent regulation, onerous capital requirements and a close-knit community of insiders.
Given the fact that many insurtechs are reliant on capital from industry incumbents, attempts at widescale disruption risk biting the hand that feeds them.
At the same time, insurers need to catch up with digitally advanced competitors. In the main, this has been achieved by acquiring a start-up or attempts at in-house innovation, using a ‘corporate lab’ – although success with the latter has been limited by the lack of internal tech expertise.
Meeting of minds
Partnership could be the best path for both parties. One move in that direction is a hybrid innovation strategy, in which an alliance is executed including equity stakes, joint-venture innovation labs and the creation of advanced data platforms for analytics, built within the insurer, benefiting from knowledge transfer within the insurtech community.
In such models, a nimble technology company retains ownership of IP and benefits from a huge wealth of data accumulation while it also builds long-standing and trusted partnerships.
The formation of the insurer’s data strategy sits neatly alongside the attempts to foster innovation and can be seen as a further driver of productivity.
A robust and transparent data strategy can balance defensive approaches including cyber security with more outward-looking objectives, such as the use of predictive analytics, under the insurtech approach of augmented intelligence.
Within the world of underwriting specifically, insurtechs are already playing a role in the evolution of specialty insurance in Lloyd’s and the London market, with AI-driven algorithms now being used to select business and provide an index-based ability to passively underwrite.
Elsewhere insurtechs are using AI to pick the best business and actively beat the index, generating ‘alpha’. The use of AI is already accelerating a division between leaders and followers in specialty insurance, which will have a profound effect on the allocation and cost of risk capital and on insurance enterprises’ earnings.
Each class of business will inevitability concentrate on a smaller group of true leaders, with followers competing using lower expense bases and diversified capital, perhaps relying on earnings from comparative leadership positions elsewhere.
The role of reinsurance
Reinsurance can play more than a supporting role in the transformation of the industry, supplying not just capital but also expertise and technology-enabled products as the catalyst for innovation.
Reinsurance could also unlock insurtech sustainability ahead of a possible convergence, with the distinctions between technology-enabled insurers and insurance-focused technology companies becoming redundant.
Reinsurers occupy a unique position at the top of the risk food-chain and along with a role in allocating capital (to insurtechs that can demonstrate adequate risk-adjusted returns), reinsurers can play a role in technology diffusion.
A convergence between venture-backed technologists and publicly listed insurers is possible, but what does that mean for the thorny issue of valuations?
My view is that premium valuations are achievable and sustainable if matched by earnings momentum and clear ownership over strategic territory, particularly when allied with a deep understanding of the client and the ability to use data to better price and manage risk.
A spectrum of valuations is still likely. Strictly regulated and balance sheet heavy insurers may benefit from improved metrics if they reduce footprint and distribution costs, but that’s unlikely to receive a premium technology rating.
On the other hand, an insurtech that can demonstrate profitable underwriting growth and durability in its capacity, including some dedicated risk capital, should be able to retain premium valuation as its IP is monetised and earnings grow.
The jury on convergence may still be out, but as pioneering author William Gibson once said: “The future is already here. It’s just not evenly distributed yet”.