One summer during college I worked in a pipe mill that made ductile iron pipe for wastewater applications. My job included painting the tapered end of each pipe, and I painted almost 100,000 pipe lengths in the stifling Alabama heat that summer. After graduation, I began working in that same company’s sales office. A few years later, I visited the production line where I had interned only to see that my workstation and the one next to it had been replaced by two painting robots.
Given this experience, my perception of automation involved a firm making capital improvements that triggered either layoffs or drastic changes in job duties. And the media attention on the relationship between automation and job loss furthers this perception. That’s why Mary Lacity and Leslie Willcocks’s ongoing, yearslong research project on robotic process automation has interested me so much. Robotic process automation challenges the sort of zero-sum thinking that I had – and many people may still have – regarding automation.
Robotic process automation works differently than the sort of automation many of us immediately imagine. RPA, unlike the automation mentioned above, uses software rather than machines to accomplish its goals. RPA efforts, as Willcocks notes, “take the robot out of the human” by using software to automate the tedious parts of a job and thus leaving the employee to do higher-level, more interesting work that requires the unique judgment and social skills we all possess. RPA thus unlocks employees’ capacity across the entire firm to do more strategic, impactful work, as most RPA automates a portion of people’s work rather than their entire job. For how RPA generates business value, Lacity and Willcocks’s case studies on Telefónica O2 and Deakin University provide valuable insight. Firms as wide ranging as the Associated Press, KPMG, Nokia, and Standard Bank have adopted RPA in their operations.
But how does RPA operate in heavier industries? What about industries, such as supply chain management, shipping, and port operations, that have experienced a great deal of volatility due to the pandemic and economic uncertainty?
A recent case study by Remko van Hoek, Jacob Gorm Larsen, and Mary Lacity provides some answers to these questions. “Robotic process automation in Maersk procurement – applicability of action principles and research opportunities,” published in the International Journal of Physical Distribution & Logistics Management, outlines the multi-year RPA adoption process underwent by Maersk’s global procurement team. In procurement, the pace of RPA adoption has been slower than in other fields, such as accounting and customer service, but the potential of RPA in supply chain strategy remains high, so long as its role is that of “an enabler and not the solution.” This potential is also complemented by the presence of a series of action principles, which “offer guidance, not rules, for the consideration and possible adoption of emerging technologies [such as RPA].”
Finding the strategy behind RPA
Maersk, while a leader in global shipping, also seeks to become “a focused logistics company with a vision to be the global integrator of container logistics.” As such, the role played by the procurement team underwent large changes – instead of being an important supporting player, it became “embedded in the business’ of the new Maersk.” Digitization and the various improvements it can offer thus became a priority, but the question of what and how to automate remained a question. In an organization as large and complex as Maersk, it’s less about whether automation will benefit the enterprise and more about what strategic goals require RPA. Yes, studies consulted and analyzed by Maersk’s procurement leadership team “indicated potential for automation of about 50% of its existing processes,” but the question remained what strategic goals automating these processes would serve.
That question of strategy is not one to overlook. After all, it is easy to become fascinated by new and emerging technologies, but a clear, defined endgame as well as a strategy to get to that endgame will ensure that an organization looking to implement RPA does so effectively. For Maersk, RPA would take tedious tasks off procurement staff members’ dockets and enable them to focus on more strategic, ambitious, and critical work, such as collaborating more deeply with suppliers and key partners to reach their goal of carbon neutrality by 2040. In short, procurement staff would have more time and energy to “focus on strategic priorities and projects.”
Knowing why RPA should be implemented and in what areas, though, is only the start of the process. Maersk procurement then had to determine what tasks would most benefit from RPA. If the tasks are repetitive – such as transferring large amounts of data from one platform to another – and thus have a high likelihood of humans becoming disinterested and making errors, then they might be a good candidate for RPA. Tasks that are simple, rules-based, and that are somewhat frequent, generally represent ideal candidates. After all, why spend time automating a report that only runs once or twice a year? For that task, simply improving the existing process may be the best option. For a report that runs daily, however, automation may make more sense.
Change management and RPA
Change may be the only constant in life, as Heraclitus argued over 2,500 years ago; the 25 millennia between Heraclitus and today, however, have demonstrated our difficulty managing change. Maersk knew the change management challenges it faced and sought to move carefully yet purposefully through the automation process. Even the name of Maersk’s robot – Holger, named after a supervisor’s pet dog – was a deliberate attempt to be “endearing and nonthreatening” to employees. Holger was generally well received, but many teams with a high automation potential had an understandably high level of anxiety. Responses ran the gamut, from fear and attempted sabotage to concerns of cost overruns to metrics for success. How, then, did Maersk make stakeholders comfortable with the myriad changes RPA adoption brought?
- Engaged stakeholders in the process – Maersk began with a bottom-up approach that allowed stakeholders to propose simple, non-critical processes “to build trust and support for Holger” before tackling larger, more strategically important candidates for RPA
- Developed stakeholder engagement by targeting tasks that impact few people first before moving onto more critical tasks. This more deliberate approach accomplishes two things: it reduces fear of change from employees, and it allows the organization to gain more experience with RPA before moving onto strategically important tasks
- Located the right metrics, then tracked and reported successes – Maersk focused on time as the key metric to track. The overall number of hours/full-time equivalent (FTE) that can be automated became the key metric to track. FTE hours are tracked in real time and reported to leadership each month. On average, Holger “performed an average of +60 FTEs, with spikes of +70 in busy weeks” in 2021
- Redesigned employee scorecards as automation expands. As RPA becomes entrenched in an organization, HR must become involved in order to retain employees; at its core, RPA is less about a technology shift than a “people change.” At Maersk, many team leads “have hard automation targets in their annual objectives”
- Stayed focused on strategic objectives and tasks that support them – RPA frees up time, and employees will find new tasks to keep them busy. Are those new tasks, though, the right tasks? Team leads and senior management should regularly assess whether these new tasks “align with organizational transformation focus areas”
Instead of just one robot, Maersk now uses eight robots and those robots work on 100+ processes. An additional 60 processes are in the pipeline and innovation has continued, as Holger has taken on more processes beyond the scope of its initial mission.
RPA and your firm
Not every industry is heavy industry, and not every supply chain firm has the resources and expertise at Maersk’s disposal. That said, every company can consider Maersk’s practice of prioritizing "strategy over technology.” RPA created a triple win by creating enterprise value, customer value, and employee value at Maersk. Additional considerations firms should think about prior to adopting or expanding RPA in their procurement operations include:
- Keeping a “human in the loop,” as Maersk calls it, by temporarily halting the automated process and having humans handle exceptions and nuanced situations before returning the process to automation. In short, companies should not completely remove humans from the equation
- Determining what the real benefit is to your organization – for Maersk, cost savings or headcount reduction were not the primary goal of RPA. As the authors demonstrate, “cost savings may only be a fraction of the benefit and possibly even the wrong objective to target RPA.” For Maersk, the real ROI wasn’t headcount reduction or productivity improvement but on procurement staff being able to focus on more critical, strategic tasks
- Focusing on gradual enhancement via RPA rather than overnight transformation may allow your firm to automate without as much pushback or anxiety from stakeholders. Automating via a more gradual, iterative process may allow your firm to develop new skills in current employees or create new roles and bring in new talent
- Having executive sponsorship for RPA efforts is key – while research is ongoing as to what stage of RPA rollout requires what level of executive sponsorship, no one doubts the importance of senior-level buy-in and their willingness to communicate to stakeholders about the triple-win benefits RPA can bring
Even if your firm isn’t considering automation, the business strategy and change management lessons provided in van Hoek, Gorm Larsen, and Lacity’s case study should not be ignored.