Gartner defines Robotic Process Automation as –
Organizations in a rush to participate in automation feel that automation is the answer to all their manual processes. Organizations starting their automation journey must pick their use cases wisely. Picking the right use cases depends on whether they can demonstrate value from such an exercise and get organizational buy-in.
What Not to Automate
Not every manual process is a good use case for automation, and so in this frenzy, it is more important to know what not to automate.
In my experience, below are the five things I think companies should keep in mind while picking a use case for the automation experiment.
- Do Not Automate a Bad Process
Before starting the automation exercise, identify what you are trying to achieve through the automation of that use case.
If the goal of the automation exercise is to go beyond labor hours saving to improving the overall process, then companies should first focus on understanding the current process and the new process that they want automation to enable before RPA implementation.
Organizations think implementing automation will improve their current processes. However, that is not the case because RPA does not redesign anything.
As Andrew Spanyi, the author of four books on process management, says: “RPA does not redesign anything. It doesn’t ask whether we need to do this activity at all. It operates at the task level and not the end-to-end process level.”
- Doing a Thorough Cost vs. Benefit Analysis to Improve Return on Investment
If tasks have a high volume of scale, and are repeatable and valuable, then they are good candidates for automation. Sometimes even if it is not hugely valuable but doing it manually makes it prone to quality error, then also it makes sense to automate unless the cost heavily outweighs the benefit.
If it is a one-time project or tasks that are low volume, then it is better not to automate.
A few years back, my team and I were helping a customer load all the excel file-based data after doing quality checks into the warehouse. The exercise took about 15 mins every month. When the team came up with an automation proposal, because of the siloed systems that were in use and technology limitations, the time and cost to build the automated solution were much higher than spending 15 mins every month.
Keep in mind the three V’s – volume, variance, and value. The activity should have a high volume and consume enough hours for delivery, low variance in the steps being used to deliver the outcome, and also comes at a high resource cost (value) to perform.
- The Process Is Complex and Needs Human Judgement
There are some powerful automation software applications in the market that are capable of handling complex processes. They can run workflows with several steps and data touchpoints, using advanced layers of “if’ rules accurately and efficiently.
If a process is complex and needs to go through multiple approval levels and decision-making may not be efficient to automate. There are also processes where the decision-making may need human expert knowledge. That is also
For e.g., A user for your enrolment dashboard asks you why the dashboard is not showing the correct enrolment numbers for their study. Though this investigation exercise may have certain simple steps like opening the dashboard and comparing the number in the dashboard with the number in the Clinical Trial management system, there are other aspects to the root cause analysis that requires someone to know how the dashboard has been built, formula applied, how the data is coming in from the CTMS system, etc.
- Involve Subject Matter Experts from the Business Side
Organizations often use the help of outside vendors to automate their business processes. The thought process here is that since the business function themselves do not have the time to do it, bringing in an outside vendor will give them the time and expertise. This would work well if automation was only a technology problem which is not always the case. Automation is predominantly a process problem. Understanding the nuances of the process should be the first step in any automation exercise, and that is why it is very important for the business function teams to be part of that mapping exercise.
Implementing an automation technology without a complete understanding of the existing process and the intended process will lead to sub-optimal value from that exercise.
- Automate Parts if not the Whole
Sometimes after doing the initial process mapping, you may realize that the whole process looks too complex to automate in total. In that case, it often helps to look at part of the processes that you can optimize through automation to reduce the overall cycle time.
Earlier in the article, I shared an example where we were helping a customer load all the excel file-based data after doing quality checks into their data warehouse. After the initial process mapping, when we realized that automating the whole process did not make sense, what we ended up doing was automating parts of the process, which reduced the processing time from 15 minutes per month to about 7 minutes.
Do not use RPA Implementation to support your current process only. Use this as an opportunity to examine the current process steps and look for improvement. Automating the current process may achieve modest savings, but you will lose out on opportunities to dramatically improve process outcomes, quality, and cycle times.
How DefineRight Can Help Decide What, When, and How to Automate
At DefineRight (DR), we understand that automation is more of a process problem vs. a technology problem. As part of our application engineering service portfolio, our focus in any automation exercise is to first map the current business process comprehensively and then recommend ways of improvement and identify automation opportunities.
