By John Warburton
Founder, Konsileo
“Robots will take over most jobs within 30 years, experts warn.”
Daily Telegraph, 13 February 2016
“According to our estimates, about 47 percent of total US employment is at risk. We further provide evidence that wages and educational attainment exhibit a strong negative relationship with an occupation’s probability of computerization.”
Carl Benedikt Frey and Michael A. Osborne
“The future of employment: How susceptible are jobs to computerisation?”
Oxford University, 2013
“If you don’t think about and plan for the future of work then your organization has no future.”
Jacob Morgan, The Future of Work, 2014
The impact of robotic process automation (RPA) and artificial intelligence (AI) on employment is becoming a matter of significant public debate. There are commentators, such as Oxford University’s Frey and Osborne, who foresee significant public policy challenges as up to 50% of roles are eliminated. There is also evidence that, especially in the transition, there is a very negative impact on wages, bringing with it further public policy and security challenges.
In some ways, there is room for pessimism. Some foresee a dystopian future where people are subordinate to either the machines themselves, or massive, coercive corporations. Consultants PwC talk about three “worlds” in the future: an “orange world” of small organizations specializing in particular markets or capabilities, a “green world” of companies with strong CSR (corporate and social responsibility) agendas; and a “blue world” of strong and (greedy) corporations. (See PwC, “The Future of Work – A Journey to 2022”.) These scenarios are convincing and challenging because we can see the trends already starting, with the end of the “job for life”, virtualization of our working lives, always-on technology, and normative changes in working culture.
“We have to accept the fact that our careers no longer go ‘up’ and we can’t depend on one company to take care of us for life.”
Josh Bersin, Forbes.com
For insurance, technology-driven change has a very mixed history. Insurance was an early adopter of technology in the 1960s and 1970s and, as a result, was one of the first industries to experience the “legacy” challenges of innovating on very complex and difficult to understand processes and systems that were created at that time. As a result, employment in the industry has remained fairly steady, at circa 2.4 m in the US1 and 305,000 in the UK2. We are, however, only part-way through what author Jeremy Rifkin has termed the “Third Industrial Revolution”. InsurTech holds the promise of propelling insurance towards what Klaus Schwab, the World Economic Forum founder, has termed the “Fourth Industrial Revolution” where physical, digital, and biological worlds are fused.
The “future of work” issue in insurance is particularly difficult to manage, not only because of legacy issues, but also because of the intangible nature of what people in insurance actually do. At its core insurance is a promise to pay money if certain things happen. The insurance industry has grown up because:
If all these tasks become more straightforward because robots/AI can handle the complex but repetitive elements, then (so the argument goes) the logic for people to be engaged in this work reduces considerably. Many believe that this wave of technology adoption will reduce employment as complexity is reduced and customers’ demands for cleaner, low-touch experiences are met by InsurTech. So, what is the future of employment in the industry and which roles will survive and prosper?
The classic framework to understand the impact of “computerization” on work is that developed by Autor et al. in 2003, which is shown in Table 1. It talks about “complementarities” between IT and various types of work; i.e. where technology enhances the effectiveness and productivity of the tasks as opposed to other tasks where technology can “substantially substitute” for labour. We can also imagine insurance tasks that will be similarly impacted. In broad terms, the routine tasks will tend to disappear and/or attract reduced wages over time while the non-routine tasks continue to attract reasonable remuneration. Since 2003, technology developments in AI will undoubtedly have moved some tasks from the right to the left-hand side of the chart below, e.g. truck driving and maybe medical diagnosis.
Table 1: Predictions of task model for the impact of computerization on four categories of workplace tasks
Routine tasks | Non-routine tasks | |
Analytic and interactive tasks | ||
General Examples |
|
|
Insurance Examples |
|
|
Computer Impact | Substantial substitution | Strong complementarities |
Manual tasks | ||
General Examples |
|
|
Insurance Examples |
|
|
Computer Impact | Substantial substitution | Limited opportunities for substitution or complementarity |
Source: After Autor et al., The skill content of recent technological change: an empirical exploration, Quarterly Journal of Economics, 2003.
Note: Insurance examples added by author
The key consideration, however, is that this analysis is conducted at a task, not a role, level. Obviously if a role exists to fulfil only one highly repeatable task within an industrialized process, then the role itself is at risk if the task is automated. This “deskilling” as a step on the road to an automation process has been a hallmark of much of the operational change in insurance, in common with nineteenth and twentieth century developments in manufacturing. Like more modern manufacturing techniques built on six-sigma and self-organizing teams, insurance organizations are now starting to wake up to the value of empowering teams to manage their own efficiency and effectiveness – this tends to prove particularly effective where tacit knowledge and empathy are most critical. This mix of tacit knowledge and empathy will continue to drive fulfilling careers in insurance.
Much of the InsurTech activity in the early wave has focused on accelerating the process of automation of routine activity. The most obvious opportunities for this are in personal lines (i.e. for consumers) where straightforward opportunities for improvement in user experience and efficiency exist. In addition, consumers in many markets have demonstrated that they are willing to self-serve. In the market for commercial insurance, the case is less clear cut and the market is not necessarily automatable, as shown in Figure 1.
There are a number of factors that both accelerate and retard the commoditization of commercial insurance underwriting, as shown in Figure 2. In essence, the desire and economic imperative of insurers to actively commoditize the market is retarded by both clients’ and brokers’ desire to retain human contact in the process as companies get larger and/or more complex.
The slowness of “e-trade” and resistance to commoditization tends to be viewed negatively, but an alternative interpretation is that market participants perceive alternative, non-rational benefits in the current trading relationships. Within this inertia may well be the seeds of a source of enduring strength in the commercial insurance market and employment within it. It suggests that empathy, longstanding relationships, and tacit knowledge have an economic value to clients. In this regard, the development of the commercial market adheres to the customer satisfaction product design principles of Noriako Kano (Attractive quality and must-be quality, Journal of the Japanese Society for Quality Control (in Japanese) 1984), much used by companies such as Amazon and which talk about:
The challenge is that much of the process innovation in the industry has tinkered with “must-be” or “one-dimensional” attributes in order to save cost. This means that many clients see changes in a negative light. Similarly, the relatively small number of “delighters” often become commonplace over time (e.g. claims tracking) and there is not a pipeline of true innovations to sustain delight. This lack of innovation in customer experience, alongside the intrinsically challenging and heterogeneous nature of commercial insurance, goes a long way to explaining the slow take-up of self-serve and new propositions in commercial insurance. Figure 3 sums up the client considerations in evaluating new propositions. Clients balance the convenience versus risk of advice together with their perception of the challenges of self-serving insurance. The risks are perceived as relatively low for small businesses but increase exponentially as businesses get larger.
The impact of technology on the careers of individual underwriters and insurance brokers means that they should:
Insurance broking and underwriting – the people who work in insurance – have every chance to become the professional service that they strive to emulate. However, a transformation in the attitude of practitioners to their role and in the support that they get from the organizations they work for will be required.