Part 1 of 2 – Learn how Process Mining can be used to identify and evaluate RPA opportunities faster and with more accuracy, than traditional methods.
I often get asked if Process Mining can “automatically identify” where RPA bots could best be deployed to automate manual processes. In today’s big data and AI world, sometimes people promise (or expect) too much.
The answer to this question is both “no” AND “yes”.
“No”, because process mining *by itself* can’t automatically identify RPA opportunities. Despite what some vendors say, it is after all, based on system event data, so manual work is likely to be invisible to it.
And “Yes”, process mining can (and should) be used to evaluate the RPA opportunities so you know which ones are the most suitable and will give the best return.
Summary
Process Mining is definitely a crucial part of the automation journey – in fact, it should be the first step in the journey. That’s because understanding your business process end to end with Process Mining is the better and more strategic approach to manage an automation programme in the long run, rather than just focusing on a small part of them.
It is important to evaluate your automation projects in a priority order based on ROI – so it is important to calculate that accurately based on facts and not just guesswork. Once the automation is implemented, it is important to monitor the entire end to end process (not just the automated piece itself) to see the impact that automation has had on the overall, as well as to see if there are unintended consequences. Furthermore, a new ability available in very few process mining tools, (which Arkturus will have very soon – watch this space!) is the feature to simulate what the outcome of automation will be, and what the downstream effects are, and if any other changes need to be made to support the change.
Process Mining can be used in the automation journey in the following four areas:
This blog is Part 1 of 2, on how to identify, and then evaluate, the automation opportunities within your business.
Step 1: Identifying and Quantifying Manual work
Identifying where a lot of manual work is done is actually the one part of process discovery that is best gained from the users. It is they, after all, who know where they are spending most of their time performing a lot of manual steps.
So, hold a workshop (whether in person or remotely) and talk to them about it – you will find Process Mining invaluable as part of this activity because having the process mined and visualised from your system data will give the discussion fact based context and allow you to quantify the automation opportunity much more efficiently.
Once these manual steps are identified, add them to the process chart that process mining gives you – there is a special function to do that where you record the manual steps or decision points.
As well as identifying where these manual areas exist, you need to capture some information that will be used as part of the evaluation criteria.
There are many articles on the internet that talk about RPA evaluation criteria, so we won’t go into them in depth here, but the main ones are:
Some of this information is already known and in the mined system data; others need qualitative input from users; others, like “Effort per case”, could benefit from accurate timings being taken.
This user derived information can be recorded against the manual step as shown here.
Step 2: Evaluating the RPA opportunities
Arkturus can help you evaluate these opportunities quickly and accurately, by combining the data you manually captured in the previous step, with data that is already known to it through the mined process. We have standard dashboard templates for this, that can be used as a starting point and customised easily for your specific organisation’s evaluation criteria, ie with the criteria that is important to your organisation heavily weighted.
Without Process Mining, you will be spending more time sourcing and analysing the system data needed to evaluate the opportunities against your criteria accurately. But with your process already mined and visualised, all the information is already there, and can be merged with the user derived information, to form a clear and fact based evaluation dashboard.
The example above shows that we have identified three manual areas through user interviews, and they are now listed as opportunities for RPA implementation.
The manual effort per case that was captured as part of discovery, combined with the volume and transaction data we already have in the mined process, gives us the cost and effort measures you can see. The standardisation measures are derived from the variations per pathway available to us in the mined data. The automation potential is derived from task inputs/outputs and the decision logic captured during discovery.
The result is a clear, effective summary of the opportunities, calculated automatically but sourced from both mined data, as well as user provided information.
Our next blog in this series explores how Process Mining can be used to simulate and monitor RPA implementations.
