Reinventing jobs: four steps to optimising work automation
HR is asking the wrong questions when it comes to the automation of jobs, says professor John Boudreau. If business wants to optimise people and machines, it needs to follow these four steps
Every day automation can do a little bit more of the work in most jobs. Every day a different kind of worker – a gig worker, a contractor – may be available to do some of the work that used to be embedded in a traditional job.
Most of the time these changes do not create disruption. However, when it reaches the point where the little more work has built up to a lot more work; where the kind of worker available to do some of the work can do more of the work – well then work itself is disrupted and people will need reskilling. And this is the fundamental dilemma and opportunity for HR, says John Boudreau, professor of management and organization and research director at the University of Southern California's Marshall School of Business and Center for Effective Organizations. How do you create organisations that are good at managing this constant upgrading of work?
“In every area of life there is the constant upgrading of technology. Most of the time you hardly notice, and you don’t have to change your behaviour. But at other times the upgrade is fundamental, and you have to change your behaviours," says Boudreau.
“Work is going to be done in many different ways. An increasing part of work is not done in the traditional employment relationship but by gig workers, volunteers, freelancers and automation. The questions are not how many jobs will be automated, in which jobs will workers be replaced by automation and where will reskilling be needed for workers who lose their jobs to automation? The questions are what work will be most affected by automation, what is human-automation collaboration now and in the future, and how can leaders and workers create a culture and practice of openness about work that is constantly upgraded?”
Take an oil rig. Today some organisations have completely automated the oil rig. There are no humans on the rig, the human is in a cockpit where he or she is remotely and digitally managing the rig. It is mental, not physical work. The person is sharing the control with artificial intelligence (AI) that can monitor better than humans and provide signals on what needs attention. But the human workers train the AI, interpret the signals and make decisions.
It’s not as simple as how many rig workers will be replaced by the automated drilling rig. Instead the question is how the human work of running a rig will change so that humans can collaborate with automation. In this case humans are doing a safer job in a more controlled and predictable environment and are higher skilled and therefore higher paid.
Boudreau and Willis Towers Watson managing director Ravin Jesuthasan discuss how business can create new work when automation opportunities present themselves in a new book, Reinventing Jobs: a 4-Step Approach for Applying Automation to Work, released by Harvard Business Publishing. They suggest a framework for business to maximise the most important opportunities to optimise jobs and work.
Step 1: Discover the automation compatibility of each task
In other words, deconstruct work to discover where automation will affect it. Start by deconstructing the work into the tasks. Only then can you find the pattern of automation and work. Once deconstructed, ask which tasks are most compatible with automation. Ask to what extent each of the tasks are likely to fall to the left or right of the below:
- Repetitive vs variable and unpredictable
- Independent vs interactive
- Physical vs mental
Step 2: Check the Return on Improved Performance for each work task
This is about understanding how automation changes the value of work. Imagine a graph with level of performance on the horizontal axis and organisational value on the vertical axis. There are four curves: negative value, constant value, incremental value, exponential value. Plot your tasks, with those in the lower left being low performance and low organisational value. Here the task is to prevent mistakes, but there is not a lot of value beyond this.
Tasks that have a slightly higher level of performance and a more positive value are referred to as constant value. Here there may be lots of ways of doing something. They all add about the same value, but there is not a lot of opportunity to add more value.
We then move to higher performance higher organisational value tasks, those that add incremental value.
Finally, on the top right are those higher performance higher organisational value tasks, the ones that create breakthrough value.
So, consider what you are automating for. Is it to prevent mistakes, add incremental value or add exponential value? What is the value created by each task and how can it improve?
Step 3: Choose the type of automation for each task
Which type of automation is best for the task:
- Robotic process automation
Used for high volume, low complexity and routine tasks. Best used where data needs to be transferred from one software system to another but requires no learning from interactions. For example, verifying customer account balances contain sufficient funds in a bank
- Cognitive automation, AI and machine learning
Used for non-routine, complex, creative and often exploratory tasks. Best used to recognise patterns and understand meaning in big data and where learning from interactions is required. For example, learning from customer data to help a bank teller better recommend additional banking services or to design new products
- Collaborative or social robotics
Used for collaborative tasks, for both routine and non-routine tasks. Best used when robots are mobile and move around among humans, are programmable and adapt to new tasks. For example, augmenting humans in a factory
Step 4: The role of automation in the work
Where will automation apply?
- Substitute for humans: automation will replace humans, for example verifying that customer account balances contain sufficient funds in a bank
- Augment humans: automation will help human workers do the same work at a higher level of performance, lower cost or greater reliability. For example, learning from customer data to help bank tellers better recommend additional banking services
- Transform the humans: automation creates new work that changes the capabilities and performance value of the human workers. For example, collaborating with bank product designers and process leaders to improve and design new products and processes
The fundamental idea is HR cannot optimise work if it only thinks about replacing workers in a job. It must deconstruct the job to see what tasks can be automated to produce the best return on improved performance.
And the time to do this is now, argues Boudreau. “If we wait until automation is available, then workers will have to reskill quickly and adjust quickly to new work. It we start early and can see automation on the horizon and create leaders that are ready to discuss changes with workers earlier enough, we will produce workers that are much more open and ready for these changes.”
The questions are what work will be most affected by automation, what is human-automation collaboration now and in the future, and how can leaders and workers create a culture and practice of openness about work that is constantly upgraded?