At what point is being good at your job no longer good enough?
A guest post by George Zarakadakis
As artificial intelligence and machine learning applications become pervasive, questions arise with regards to rationalising jobs and managing human talent. Businesses must decide how best to balance the efficiencies of automating tasks using intelligent machines while at the same time managing the risks of automating too quickly or slowly. A useful tool in deciding how jobs may be affected in the era of robotics and cognitive machines is the ‘Return on Improved Performance’ (ROIP) curve shown in the graph. Using it, companies can make the important distinction between two categories of role: ‘proficiency’ and ‘pivotal’.
Let’s take the airline industry as an example. A typical ‘proficiency’ role is that of a pilot, whose performance increases as they notch up flying hours. They achieve a minimum performance level – when they get their licence – then go up a level when they become a captain.
For as long as they are progressing their careers towards becoming captains, each pilot keeps returning higher value to the organisation. However, once they have achieved that level, the value returned ‘flattens out’ since they will continue to perform essentially the same tasks – commanding the airplane, supervising the junior crew – while continuing to demand more pay and benefits due to their longer years of service.
When good just isn’t enough
Since all proficiency roles have a similar relationship between performance and value, they are highly susceptible to transformative disruption by intelligent machines. It is not impossible to imagine a future in which pilots have been replaced by automation since the latest generation of commercial aircraft is already capable of autonomous take off and landing. While most passengers might currently feel safer knowing that human pilots are sitting up front, there is also an increasing awareness that the majority of crashes are caused by pilot error and that computers should be given the ability to override human decisions if the safety of the plane is at stake. If that sounds too much like HAL from 2001: A Space Odyssey, consider the recent high profile cases of planes crashing either due to poor pilot judgement or personal psychological issues.
Pivoting on the human touch
If a pilot is an example of a ‘proficiency’ role, let’s now turn to the flight attendant as a ‘pivotal’ position. In this job, great talent really can be a differentiator since an ever-increasing performance level can keep returning more and more value to the organisation. A talented flight attendant can deliver superior customer satisfaction and with that, lasting customer loyalty and advocacy.
Pivotal roles are hard to replace by AI since creativity and the human touch are hard to define and harder to replicate. However, they can be susceptible to augmentative disruption by intelligent machines. Imagine the benefits of a flight attendant using a Google Glass type device to access information on each passenger in real time. Equipped with such machine learning-powered technology, the flight attendant can augment their performance and deliver even greater value to passengers and, therefore, to the organisation.
Retain or retrain?
Over the coming decade, companies will need to be able to identify and segment jobs according to the pivotal or proficiency characteristics in order to make the right decisions about automation. Pivotal jobs will be the clear winners in the fourth industrial revolution but what about proficiency jobs? Will they be eradicated across sectors, the way pilots stand to become substituted by fully-autonomous planes? Without doubt, some proficiency jobs will go but for most roles, automation will offer opportunities for them to become transformed into pivotal jobs. Companies will need to analyse these opportunities in order to retain vital talent while whitecollar workers will need to assess their skill sets and retrain, as well as retool, for a future where the ability to manage human interactions will become more valuable than task-based skills, since it will be the hardest thing to automate.
Dr George Zarkadakis is digital lead at Willis Towers Watson