Many companies are tentative about how to react and where to employ intelligent machines into their daily operations. Some large insurance and brokerage firms have begun dabbling with robotic process automation (RPA) as a way to reduce costs by automating outsourced and low-risk work.
While the intentions behind them are sound, many RPA initiatives have failed to achieve the desired cost-containment objectives, however. This is partly due to “overpromising and under-delivering” supply, but also because early buyers don’t yet know where to apply the promising technology to the most suitable business cases.
While RPA is most commonly associated with the current state of automation, it is important to understand that RPA is merely the most basic type of such technology. It addresses rote, repetitive or otherwise routine work that can be easily programmed to or carried out by a computer.
Cognitive or smart automation, on the other hand, involves more complex activities, such as natural language processing, artificial intelligence (AI), sentiment analysis and self-learning. At this level, machines can work independently and without human instruction or with only periodic human intervention and/or approval. While the former type is well under way, the latter is only in the nascent stages.
Before an organization can determine which type of automation its culture and processes are suitable, it must first conduct a cost-benefit and maturity analysis to help focus on processes that make the most business sense. Of course, some business processes need not be considered for automation. For example, it may be fruitless for an organization to incur the costs of automating an activity that is performed only once a year or has an insignificant effect on volume or resource allocation.
Setting the score
With the help of an automation scorecard provided by us or other integrator, the first phase of assessment breaks down all processes and subprocesses that are being considered for automation. At the end of this step, an enterprise-wide functional inventory should be listed and well understood.
The next step includes interviews with business process owners and stakeholders to ensure the most accurate and useful information is captured for further analysis. For example, underwriting as a function is composed of both the individual underwriter and the actual process of underwriting. Hence, we initiate the identification phase by choosing parameters that affect both the individual and the business process: the degree of work variation (including project management and analytic skills required) and the degree of business process complexity (including both the process traits and business impact).
Once an automation score is acquired, individual organizations or business units can begin piloting, phased development and eventual application of cognitive automation in addition to early RPA deployments. Obviously, selection of the right processes and the right level of automation can be tricky. To overcome this oversight, it’s important to get the right stakeholders committed to the program early in the game.
Smart machines, or those endowed with AI, have been described by the World Economic Forum as the fourth wave of the industrial revolution (after steam power in 1784, mass production in 1870, and electronic computing in 1969). Before planning your approach to catching this wave, it’s important to note that the automation adoption curve is extremely steep, no industry will go untouched and, like previous industrial waves, this will fundamentally change the way we all work.
Moving forward, insurers must view things from a dual lens: A microscopic view for achieving the low-hanging fruits of RPA, and a telescopic view for cognitive or smart automation opportunities in the not-so-distant future. A detailed plan of action should balance these two goals, because today’s successful RPA initiatives can provide a foundation for tomorrow’s cognitive automation. It will also help win the trust of senior management by demonstrating scalable value.
Also bear in mind that smart machines will coexist with humans in the workplace in the near future. Therefore, organizations should be proactive in smoothing the transition.