As organizations start down paths toward deploying digital process automation, it’s always helpful (and wise) to learn from other companies that have already started their journeys. Here are five valuable lessons learned from companies and government agencies that have implemented digital process automation, ranging from smaller business-led, low-code initiatives to larger, low-code business/IT collaborative projects:
1. Set clear business process goals without taking on too much scope
This statement amounts to motherhood-and-apple-pie, but in digital process automation, this simple truism is absolutely necessary. Why? All the pundits imply that process automation projects are a failure unless they transform the business from the outside-in, deliver unparalleled customer experiences, and radically shift the organization’s competitive goalposts. These lofty goals are admirable and achievable, but only 20-30% of process automation projects have that mandatory C-suite orientation for end-to-end transformational processes.
The majority of projects are instead targeted at business operations, usually with a smaller scope and a clearer bead on increasing productivity, reducing costs, eliminating waste, shifting the workforce to higher-value work, lowering rework, and better compliance. If processes in business operations are broken or need improvement, that’s often the best place to start. In fact, beginning with smaller projects for early successes is a winning strategy that organizations can build upon over time.
2. Look for processes with lots of hand-offs between groups
Sometimes project teams know exactly what process they want to start with, but often they have choices about which process(es) to launch first. If that’s the case, experienced managers recommend looking for processes with many hand-offs between workgroups, departments, or even organizations. Whenever this happens, there are high chances the hand-offs will be delayed, error-prone, or lost forever—especially if the handoffs involve manual (e.g., paper forms) or semi-automated (e.g., e-mails, phone calls, calendar reminders) approaches.
Plus, the possibility of mistakes increases many-fold when hand-offs occur between external individuals and organizations, such as customers (e.g., turnaround documents that need signing) or business partners that complete parts of the process before returning work to the originator. By targeting these situations, project teams will most likely find truly broken processes and opportunities for high ROI through automation. And it’s a funny thing—we often get so accustomed to the brokenness in a process that sometimes we forget how mistake-laden these hand-offs can be.
3. Embrace Agile and stay away from Waterfall
After all, that’s been written and said about the dangers of waterfall methodology, you’d think project teams would know about the problems. But still, many groups—especially teams with little IT involvement—launch process automation projects using a classic waterfall approach for systems development. The biggest problem is that waterfall is sequential, inflexible, and downright rigid. It’s very hard to use waterfall when business process improvements are unclear or if the processes are highly dynamic, as in case management.
In many ways, waterfall is antithetical to digital process automation, which is designed to handle iterative process discovery and process design, and support highly dynamic work. Experienced project teams warn about the dangers of waterfall and give a big thumbs up to training project team members in Agile and working with IT to use Agile most effectively on process automation projects.
4. Focus on data first
Even though digital process automation is all about business processes, experienced project managers agree that data is the hardest part to crack in automation projects. Why? Successful large-scale projects need a strong data model and require developers and the data team to work closely together on data virtualization. According to one IT executive, “My advice . . . don’t start with the process; start with the data first.
We had dirty data in our legacy system and wish we had brought data architects on board 9-12 months before starting our process automation work.” This executive advises other organizations to build a strong data architecture for the “as is” and “to be” processes before doing process modeling. Two other reasons: 1) data will become increasingly critical as organizations begin to implement AI and machine learning, and 2) focusing on data consistency and reuse is critically important for organizations that plan to automate many business processes over time.
5. Get a clear understanding of RPA and digital process automation
It’s fair to say that robotic process automation technology has taken the business and technology worlds by storm, leading to widespread sales and sky-high valuations for RPA vendors. Amazingly, with some simple screen scraping and scripting—business people can quickly develop robots that tackle massive amounts of drudge work in short order. Whether it’s cutting and pasting information from an email into forms, transferring data from spreadsheets into business applications, or handling chat sessions, these robots look like the wave of the future.
However, many organizations have confused digital process automation software—which automates processes— with RPA, which tackles tasks. In other words, RPA cannot substitute for digital process automation. Instead, RPA can automate specific tasks (including very high volumes of previously manual work, now done in minutes) that are part of a bigger process, which is frequently automated by digital process automation. The two technologies work well together; RPA is not a substitute for digital process automation.
These best practices are important for teams to keep in mind, whether the focus is on smaller projects or on large-scale, end-to-end process transformation. Here’s the shorthand advice: 1) consider tackling smaller-scoped projects initially, 2) look for broken processes that involve hand-offs, 3) embrace Agile, 4) keep your eyes focused on data as well as process, and 5) understand the difference between RPA and digital process automation.