You can increase the value of Robotic Process Automation (RPA) significantly if you start by understanding and optimizing your processes. In this blog post, you can read more about the 4 steps of successful RPA implementation using Process Mining:
1. Understand the current state of your processes
2. Streamline and optimize processes
3. Identify where your bots should be implemented
4. Monitor automation rates, process compliance, and other KPIs
Editor's note: This post was originally published in July 2017, and has been updated for accuracy and comprehensiveness by Saara Bergman in July 2020.
What is RPA?
For the past couple of years, Robotic Process Automation (RPA) has been a hot topic among business process experts. Many companies have begun to look into RPA as a tool to improve their operations, but what exactly does it bring to the table?
Robotic Process Automation offers an approach for automating manual tasks - providing a way to get high-volume tasks performed fast and without errors. The focus of RPA is to carry out these tasks automatically on the existing software front-end.
RPA vs. Traditional Workflow Automation
Traditional workflow automation operates with structured data and uses data integration methods and heavy scripting to achieve its goals. RPA focuses on automating tasks that handle unstructured data while aiming to operate on the same level as an end user. The results include streamlined repetitive high-volume processes, automated manual work, and the relocation of experts’ time for more profitable tasks.
RPA makes operations accurate, as software robots repeat the task with pinpoint accuracy and efficiency, and are scalable in accordance with the demand. All these actions produce data useful for analyzing performance. However, Gartner points out that while RPA has a precise definition, the offerings on the market are varied and do not necessarily represent the proper definition.
What is Process Mining?
Process mining refers to the method of analyzing data gathered from information systems.
As information systems produce and store huge amounts of data as logs, a lot of unused data lies dormant. This data contains valuable information on how each step of the process is carried out. The access to this valuable information creates a venue for process mining tools to show their worth.
Read: What is Process Mining?
How does Process Mining support RPA implementation?
By using modern data mining methods, the unused process data is made useful. Process mining software brings you this insight by drawing data directly from your IT systems (such as ERP, CRM and BPM systems) with built-in connectors.
Process mining software, like QPR ProcessAnalyzer, then visualizes and analyzes process flows that take place in your organization, providing a holistic view of processes and identifying bottlenecks, variations, and underlying root causes of inefficiencies. This enables you to take control of the process by clarifying points that need to be improved.
Moreover, it helps you monitor automation rates, spot automation opportunities, and much more. QPR ProcessAnalyzer can be used together with any RPA software.
Successful RPA projects enabled by Process Mining in 4 steps
Although RPA has indeed been considered a silver lining for many organizations' digital transformation initiatives, reports from leading consulting firms state that half of the RPA projects still fail in getting the desired ROI. This is explained by the complexity of the actual processes.The fact is that RPA alone cannot be used to automate full, complex, end-to-end processes.
RPA works best on process steps with few variations and little complexity. Therefore, in order to succeed in RPA, you need to be aware of the state of your processes, optimize them, identify automation opportunities, and continuously monitor your bots.
1. Understand the current state of your processes
“Improvement cannot be done without understanding the current state”. This applies to process improvement – how could you improve process efficiency with robotics without knowledge about what the current process state is or what kinds of deviations or variations exist? Process mining brings vital insight by revealing the as-is process state from a data-driven perspective.
Identify long process durations, bottlenecks, and processes with high amounts of variations
Process Mining helps you check for instance where there are simple process steps, with few variations yet large business volumes, that you will gain a lot from by automating. To ensure your processes are RPA-ready, you need other insights from process mining as well.
Start by discovering your current processes: view the automatically generated, real-time process flowcharts and check for long durations, process bottlenecks, and complex processes with high amounts of variations.
2. Streamline and optimize processes
The most common reason for running into unforeseen problems in RPA is that your processes have too many exceptions. While these process variations are inevitable, it is neither practical nor profitable to use RPA on all process variations - but only on the most common ones. The more variations that robots need to cope with, the higher the costs of RPA as well as the risk of running into unforeseen issues.
Think of it in this way: If all processes behaved the way you think, you would not be running into unpredictable problems. However, it's the 20 % of the processes that do not behave the way you think that are causing 80 % of the costs and the effort for you. These processes are the one's that you need to improve before implementing your bots. If you put a robot into a process without knowing how many exceptions exist in this process, your RPA project is very likely to fail.
Consequently, creating high volumes per variation will ensure that the automation will be as beneficial as possible. Process Mining allows you to see how many cases of exceptions you have in your processes and the root causes for these exceptions. You can use Process Mining to decrease the number of exceptions, as well as to pinpoint the parts where you can automate easily.
3. Identify where RPA should be implemented
You should automate processes with regular bottlenecks, process steps with human errors, and
activities that staff hates doing (source: Gartner). Automation opportunities include:
- Your regular bottlenecks - the process steps that are jamming the whole process.
- Tasks in which your employees usually make mistakes - where manual
work is much less efficient than automated. The Process Mining algorithms
make it easy to find the root causes of these mistakes.
- Tasks that are disliked by employees because they are too dull, repetitive,
or too demanding.
You can also see these in the "Automation Opportunity Scout", which is part of QPR ProcessAnalyzer. In this chart, you should look for:
- Big bubbles, which indicate a lot of manual events
- Position in the vertical axis - event types in the 100% line are completely
manual, event types in between are partially manual/automated
- Position in the horizontal axis - high volume event types are on the right
Implementing RPA is then a pure RPA project step of constructing the rules for robots and programming the workflow execution.
4. Monitor automation rates, process compliance, and other KPIs continuously
The results of automation projects can be challenging to measure. Therefore, process mining is essential.
By continuously measuring and ensuring process compliance, you will be better equipped to establish long-term transformation and increase the use of RPA to span as many processes as possible.
With process mining, you can review that RPA achieved the desired outcome, and that processes are executed as designed. This kind of process benchmarking and process compliance measuring is an effective way to focus your investments more accurately and with a higher return on investment.
Monitor for instance the automation rate trend and touchless cases:
- Manual Activities per Case is an easy-to-use way of showing the amount of manual activities needed to complete a case. Touchless cases are cases that complete automatically, without any manual activities. Other cases may need one, two, three, or even more manual user activities in order to complete.
- The Automation Rate Trend shows the progress and results of your RPA and automation efforts as a monthly trend. It is easy to see how the automation rate is increasing as a result of RPA activities.
However, if the automation is not working as it should, the amount of other manual tasks within the process starts to increase. This is often caused by rework, manual work for fixing events performed by
bots, and manual work based on customer complaints.
All in all,
RPA is an effective automation method - when used correctly. When the process state is unknown or incorrect pre-implementation, processes are being automated, but the results are neither effective nor profitable. Therefore, process mining should be used side by side with RPA projects, to guarantee a high return on investments.
Feel free to check out QPR's process mining solution for RPA.