As I watch client after client move toward the promise of Robotic Process Automation (RPA) as a solution to their problems, I am amazed at their disregard for the small print.
For over a hundred years, organizations have been automating processes. Shortly after the first process was automated, it was noted that automation does not improve a process, it simply gives you more consistent errors in less time. Although this realization seems obvious when you sit back and think about it, many organizations believe that launching into automation without process optimization will solve their problems.
Beyond common sense, a quick web search will reveal article after article that support process improvement before automation. From the Harvard Business Review to one of the leading RPA companies, industry thought leaders are documenting issues related to implementation before improvement.
So, how do we change the current dynamic?
- Get Serious about Systems Thinking: Both IT experts and process experts claim to be advocates of Systems Thinking, yet both fall into one dimensional thinking as soon as projects start. RPA is a blending of process and technology. It is time to truly leverage systems thinking in design, deployment, and operation of technology.
- Integrate Existing Capabilities: Integrate Agency Performance Improvement Officer (PIO), Continuous Process Improvement (CPI), and Software Development capabilities into a process automation center of excellence that manages integrated process automation and management work cells. In doing so, ensure an agile approach is developed that borrows relevant methods from Lean to ensure alignment with value and detection of waste and with Six Sigma to ensure processes are properly measured, controlled, and continuously improved.
- Get Leadership to Accept Transparency: Agency leadership must embrace transparency, understanding that everyone is going to show both good and bad numbers. Leadership must accept that numbers showing poor performance are not an indictment; they are an opportunity to improve. Numbers that are always in the green mean they are probably not pushing the envelope hard enough. At the macro level, an enterprise can be connected at the key stroke level creating an opportunity for performance measurement like never before, but that will not happen if Leadership is afraid of the numbers.
- Get Real about Risk Management: With great power comes great responsibility. RPA presents an opportunity for incredible breakthroughs in operational performance and the opportunity to allow humans to focus on innovation and other value adding activities. The risk is that the robots make errors, very fast, causing serious problems. Agencies must implement industrial grade risk analysis and management to overcome leadership concerns and to mitigate the risks of serious failure in deployed bots.
Possibly unlike any technology trend that has come before, RPA presents a critical need for process improvement prior to automation. As eluded to above, there are numerous risks inherit to having intelligent software conduct transactions in place of humans. Consider that most tasks completed by humans generate some degree of defects, but they generate them at a pace humans can detect and remediate. With RPA, these mistakes will be generated and a pace imperceptible by humans. Imagine if these mistakes are things like amounts on invoices or late fee calculations. Massive numbers of these mistakes will accumulate without detection costing organizations millions in losses and litigation. Alternatively, RPA projects could and should use proven process engineering methods such as Lean and/or Six Sigma to first improve processes into an effective model the generates desired outcomes with quantifiable external and internal performance metrics. An improved process defined in this way will ensure alignment with organizational value streams and allow instrumentation and reporting of procedural performance. Ideally, proven process monitoring techniques and tools will be used to automatically analyze the massive amount of data generated and hence identify either negative trends or defects before they are out of control.
For those familiar with the use of tools such as Lean and Six Sigma in conjunction with process automation, you know that this approach is not only effective, it is actually easy and usually leads to a reduce project time line from idea to full deployment. Sure, one can get to an initial deployment by avoiding process improvement and diving straight into automating processes as they are defined by stakeholders, but the cost is multiple iterations of poor performing releases extending the achievement of a stable and fully deployed solution.
So, if you are looking at Robotic Process Automation as a solution, or your RPA efforts are not achieving the results that you expected or were promised, ask yourself if you have an integrated solution and if the problem is the processes or your program. Remember, there is never a silver bullet and that stovepiped approaches rarely work.