The Evolving Optimization vs. Automation Debate
Jun 21, 2019
Organizations exploring the use of Robotic Process Automation (RPA) tools to automate business processes face a basic dilemma:
- Should we reengineer existing ways of doing things before applying RPA?
- Or should we simply automate processes the way they are?
A compelling case can be made for either approach.
Proponents of pre-automation reengineering say optimization is essential to achieving full value. The most basic argument is that automating poorly designed processes will only reduce the cost of doing things poorly. And while applying automation to existing processes will almost certainly reduce costs and improve productivity, this approach fails to seize opportunities to improve process outcomes, quality and cycle times.
One major reason to optimize first is that existing, people-based processes are often inordinately complex, which means developers must invest time and effort automating unnecessary steps. With people-based processes, moreover, the rules surrounding different steps in a given process are often ambiguous; this ambiguity presents a challenge for RPA tools that require clearly defined if/then rules to execute tasks.
Another issue with automating existing processes is a missed opportunity to add value elsewhere in the supply chain. In other words, by optimizing processes, you can allow robots to focus on the repetitive administrative tasks, and allow people to apply their judgement and business expertise.
Automate First, Optimize Later
On the other side of the fence, automation-firsters argue that applying RPA to existing ways of doing things can be viable and deliver significant and rapid ROI. First off, when it comes to automating a convoluted, multi-step process, robots don’t care how many steps are involved. Moreover, while people will invariably make mistakes or take shortcuts when confronted with complex workflows, robots will strictly adhere to the rules they’ve been taught.
This “paving the cowpath” approach can also be less risky. While existing processes may be inefficient, they are at least established. Implementing automation while designing new processes in parallel, meanwhile, can introduce new and unforeseen glitches. Business expectations are another consideration. Automation-firsters say that their approach delivers quick ROI and builds momentum for additional investment, while optimizers bog their projects down in endless dithering and process tweaking.
The Big Picture
Ultimately, the optimization vs. automation debate is more nuanced than it might appear. Both sides agree that automating clearly broken processes is counterproductive. They also agree that delivering business benefit must ultimately prevail over process perfections. And while many early initiatives created discrete islands of automation, there’s growing consensus that RPA requires CIO oversight and must be part of a broader, enterprise-wide strategy.
Whatever the specifics of the optimization/automation balance, an effective operational strategy seeks to deliver significant short-term benefits while progressing towards a clearly defined long-term strategy. The challenge lies in reconciling the reality of the existing environment with the clean slate vision of the desired future state
In a recent article in Future of Sourcing magazine, my colleague Mark Popolano discusses the operational “ghosts” of entrenched systems and processes, and how these legacies limit the benefits of automation initiatives. Specifically, he writes, “the benefits [of RPA] are measured and defined in the context of people- and paper-based operations. RPA benefits of productivity, auditability and cost savings, for example, are typically gauged and expressed relative to human administrators. And, RPA sales pitches often boast that bots execute routine administrative tasks in much the same way that people do.”
The problem with this, according to Mark, is that existing processes, once optimized, will still be largely “designed to accommodate the functionality of manual typewriters and carbon copies.” What we need, he says, is to “create entirely new operational designs, process workflows and job functions. And in the process, we need to redefine the relationship between human and digital labor.”
In this broader context, the optimization/automation question goes beyond how best to improve specific processes and workflows. Instead, the focus needs to be on how to evolve and transition to a fundamentally new way of doing business.