A common challenge to both justify and sustain a successful Robotic Process Automation (RPA) program is in creating a strong pipeline of the best process candidates. I understand that not every organization has the desire nor capacity to build out a self-sustaining RPA CoE that will scale. In some cases, the benefits of RPA are fairly limited and the cost cannot be justified. With RPA now adopted for Cloud, there are many methods to implement RPA without taking on a full cost of a CoE.
So, how do you find processes for the pipeline?
To find the best processes for the pipeline, you can use a Business Process Discovery (BPD) method. BPD is the discovery of processes within the business utilizing methods to define, analyze and measure the process with the intent to optimize and/or automate. There are two approaches to BPD, Manual and Automated.
Manual process discovery is relatively low cost and engages the rest of the organization in your RPA journey. It's a great way to get them excited by creating contests and giving recognition to those that are valuable contributors. It also provides a new career progression path - from back-office agent - to RPA Analyst, or even RPA Developer. This level of engagement and new opportunities are important to your organizational culture.
The more traditional, manual approach requires analysts and Subject Matter Experts (SMEs) to work alongside each other, engaging in workshops and interviews as they discover and document the details of the process.
The other approach, Automated Process Discovery (APD), has received greater industry recognition in recent years. It promises to uncover the reality of a wider set of business processes - at scale - without the high resource demand on Stakeholders or SMEs; so much that it is even finding it's way into RPA products such as Automation Anywhere Discovery Bot, Kryon Process Discovery and UiPath Process Mining & Task Mining.
Process mining vs task mining
There are two techniques of Automated Process Discovery, Process Mining and Task Mining, each has its own merits.
Process Mining is a set of methods and tools that analyze system event logs to produce factual data on sequences of tasks that are being performed within the system. Analysts then model or organize sequences into processes for further analysis. Some Process Mining tools have pre-defined processes for specific target systems which can enable faster results.
Task Mining produces a similar output to Process Mining, however, the data is sourced from users' desktop activity. We're talking about installing software on desktops to record user actions and then centrally storing data of multiple users for analysis. Yep, your Security and IT leads just got a strange twitch as you read that sentence.
Recording users' desktop activity is a fairly invasive solution, so much that it will require a high level of scrutiny. This activity requires additional time to configure and validate ensuring that any PII, other secure information and personal activity are not recorded. Once the data is collected additional effort is needed to review the sequences of captured tasks and organize them into processes. Unlike process mining, there are no predefined processes.
Process Mining presents you with the backend transactions while Task Mining provides you with a view of user desktop activity across multiple applications. Both Process Mining and Task Mining generate petri-net like diagrams showing you the flow of volume throughout related tasks, order of actions and variants within these flows. These two methods generate different intelligence and each requires their own steps to produce usable output. Process Mining can be performed continuously and monitor processes moving forward, whereas Task Mining is generally not implemented as a continuous process improvement solution.
Accelerating RPA with task mining
When expanding automation to other business areas, we often hear that the business is too busy doing the work to spend time making the process better. A task mining tool is a potential remedy to this problem and can produce good results to accelerate RPA.
With respect to RPA, some of Task Mining’s best features are the ability to extract keystroke level actions from its recorded users, reveal variants in a process, generate actual metrics and the ability output screenshots of a defined process that are required to build a process definition document (PDD).
A PDD must reveal exactly what the robot should be doing, step-by-step, so it shouldn’t be based solely upon screenshots. Additional knowledge of the process is still necessary, such as understanding the business decisions being made, timing and other constraints. The analyst is still required to work with the SME to flush out the full details before completing the PDD.
This sounds great! I'm all-in. Before you take your first step, much like when a process is assessed for ROI, the same must be done for the task mining tool, so here are some key considerations:
- Gain acceptance of this type of tool from security and IT teams, and in terms of change management, from HR too
- Configure for any redaction and personal transactions
- Allow time to analyze the results which require SME support to help identify what tasks belong to which process
- Allow analyst and SME time to complete a PDD
Clearly, with task mining there’s less interruption to the business users, you get to see actual metric data and screenshots of what they are doing. This approach does save some steps, delivering more accurate metrics and a window to the reality of a process. However, it doesn’t completely avoid the original challenge of spending time attaining knowledge of a process with SMEs.
Going for gold
Process mining isn’t really about delivering process discovery benefits for an RPA program, but if your organization already has such a tool, take advantage of its process management and improvement analysis features to find automation opportunities. Task mining, while being very beneficial for outlining task-sequences and their metrics, will not auto-generate a completed PDD. It will provide assistance in getting you there faster and provide wider coverage of processes.
Ultimately, you must assess the time/cost benefits of using any task mining solution. Consider how well it scales and how long you’d need to retain it to complete your assessments - versus the resources required to get the same capability and output. Regardless of what approach you take, you still need to know your business. The tools can't tell you where to start. Do your diligence and ensure that you have a good plan, and above all, make sure that you can demonstrate value.
Damian Gomes, Head of Delivery, Robiquity