There can, unfortunately, be a belief among executives that automation is a “silver bullet” that can solve numerous problems around process and cost containment. There’s some evidence that some jobs lost to COVID are going to RPA or Robotic Process Automation. So, overall, the idea of automation can seem like a business concept that can swoop in and solve multiple issues at once. That’s not entirely true. While RPA can be very effective, there are specific processes that are better automated than others. That’s one place to begin your journey without even considering automation.
What are those processes?
- Highly manual and repetitive processes: Think about high-transaction volume or processes that occur daily. Accounts receivable and invoicing fall here, often.
- Prone to human error: What are some of your processes most prone to human error, that then causes additional work for others? Some HR functionality typically applies here.
- Rule-based processes: Think of processes that follow a specific set of “if-then” type rules, and can easily be turned into a template. When decision-making is based on a series of standardized rules, that process makes sense as an automation contender.
- Low exception rate: These would be processes with a low number of various scenarios. Exceptions create complexity, and a project can often be completed by partially excluding exceptions with low volumes.
- Standard readable electronic input type: Excel, Word, XML, readable PDFs. There is now RPA work being done with unreadable types, including OCR — but that does require an assessment of what’s necessary to make it readable.
- High volumes: high-transaction volume and high frequency.
- Mature and stable processes: Think of well-documented processes with known operational costs. Again, billing often falls into this bucket. Vendor selection RFP could as well. These are some of the minimum criteria to begin thinking about an RPA project.
There are a few other things to take into account early in an RPA journey, though:
- Automation savings: It is recommended that you only automate processes that would yield savings. Cost containment is a major factor in RPA selection. It’s not necessarily about reducing jobs, though, as much as letting human beings work on valuable, critical tasks and not waste time on the repetitive processes.
- Short-term changes: If there’s a high likelihood that a process might change in the short term, we do not recommend automating it. It will lead to webs of confusion, potentially.
- Go-Live expectations: For processes with relatively low complexity, it can take as little as three weeks to go live with a bot to aid in the process. Higher complexity is likely to be nine weeks or more, although it does vary by industry, vertical, and process itself.
We will have much more on automation and RPA coming in the next few months. This is but a start. Do any projects of yours come to mind as an RPA potential? If you want to have a discussion about what this would look like internally or for your end-users, we’d be happy to.