Approach for Effective Process Discovery on RPA Projects

Lahiru Fernando
10 min readJun 28, 2020


Era of Automation | Issue 04 | 31-Jun-2020

Robotic Process Automation (RPA) is one of the booming industries in the current business world. Especially with the COVID-19 effects, more and more companies ranging from large to small scale are exploring ways to get their routine, highly manual processes automated to keep the business floating with minimal human supervision. While more and more organizations start their RPA journey, a question remains on how best they know to capture and understand how and what processes to be automated. The technology companies who onboard with their RPA journey also may find it challenging to decide what is the best way to go about through the requirement gathering phase of RPA projects.

Top challenges in adopting RPA

According to Global RPA Survey conducted by Protiviti in the year 2019, the biggest challenge is understanding and prioritizing the automation initiatives. Similar to any software project, requirement gathering is a critical phase in the RPA project life cycle. Requirements for the automation should be accurately captured at a very detailed level to decide what is the best approach to take for implementation. We, as professionals in the industry, have seen companies onboarding with RPA, but not gaining many benefits through their automation initiatives. One of the main reasons is not being able to capture and understand the proper requirements.

Some believe that gathering the requirements for an RPA project is similar to the requirement gathering sessions of a software development project. It is one of the biggest mistakes most people do during the requirement gathering sessions, and later go on with so many changes in the design affecting the project timelines and increasing the costs of the project.

The best approach to understand the automation requirement

In scenarios where the organization is new to RPA, and do not have much experience in prioritizing or identifying the automation initiatives, the focus should be on process discovery to understand what are the best processes to automate to gain the highest benefit. If the organization has expertise in conducting process discovery and has a specific requirement, the best method is first to make an end-to-end detail process discovery to understand the overall picture of the process as the first step.

Approach for Process Discovery

The companies new to the automation journey could first identify the potential candidate processes for automation. What are the best candidate processes for automation? Those can be determined based on the following criteria.

  • Rule-based — The process should consist of a strict set of pre-defined rules where all employees follow to complete the task. Considering the business processes laid down in organizations, about 70–80% of these processes are rule-based and are suitable for automation through Robotic Process Automation (RPA) tools. Some of these processes might contain some level of decision making. However, the decisions backed by the data used in the process can be fed into the robots so that bots can perform the decisions accurately and efficiently. Yet, if any such human involvement is needed, those scenarios can be broken down into smaller parts and considered separately because, with the latest technological advancements, bots are quite intelligent in using humans in the loop to handle complex decision-making scenarios.
  • Standardized Processes — The process should be well established and shouldn’t change frequency. These changes include changes in the process steps, changes in the process execution order, or changes in the systems involved. The business analysts who are involved in analyzing the processes should consider these factors when deciding whether the process is suitable for automation. Further, highly manual processes might have different ways of achieving the same result. For example, two employees handling the invoice processing of a retail organization might have different steps of getting the invoices into the system and verified depending on their familiarity with the process. Such processes should go through proper Process Standardization mechanisms to optimize and standardize the process itself before applying automation on top of it. It is one crucial aspect that many try to avoid due to reasons like the cost and time taken to complete the automation of the process. However, comparing the benefit of automation over the value made, it provides much more significant benefits such as improved efficiency: reliability and increased ROI in automating a standardized process.
  • Structured Input Data — Make sure the process uses a pre-structured, and digital form of data to carry out the steps in the process. Some examples of such input data are data from excel sheets, databases, JSON files, XML files, HTML files, and many more. It does not mean that robots cannot handle unstructured data. The reason behind to look for structured data is that it is pretty much more comfortable with automating processes due to its less complicated nature. On the other hand, the robots can also process unstructured data such as data in scanned documents, handwritten text, emails by introducing artificial intelligence, OCR Optical Character Recognition (OCR), Intelligent Character Recognition (ICR) and machine learning capabilities.
  • Repetitive — The processes performed regularly are a better fit for automation as those will generate more significant positive results than the business processes that execute occasionally.
  • Time Consuming and Critical — Processes that take a long time to complete due to high transaction volume are a better fit for automation.
  • Low Exception Rate — How often the process run into exceptions? Are there exceptional cases reported regularly? Such business processes are not a good fit for automation as those processes include too many/ or even unknown business exceptions. Such business processes should be analyzed to check whether there is any possibility of standardizing them before applying RPA.
  • Prone to Mistakes — Processes that are prone to mistakes are a good fit for automation as robots could perform the same set of steps with a higher level of accuracy that can eliminate errors and improve the efficiency and reliability of the process and the data.
  • Multiple Systems Involved — Processes that communicate with many disintegrate systems are a good fit for automation as software robots can easily communicate with any software application and perform the same set of actions that a human would do without any human involvement.

As we discussed the criteria to identify the best candidate process to automate, the next question is how to identify these candidate processes? There are many tools provided by UiPath to support this exact requirement. As we know, going through such complex operations and analyzing them in detail, understanding the data and the systems, and analyzing the result to find the best process require a lot of human effort, expertise, and time. To address this scenario, UiPath is offering multiple process discovery tools to perform process mining, analyzing as well as to document the steps carried out in the process.

UiPath Process Mining uses the already available data in the existing applications to get a thorough understanding of the process to find the best suitable business process to automate, and how to do the automation to achieve the best results. Process Mining generates a high-level view of the entire process using the data in the enterprise applications. The analysis creates high-level and detail level workflows, which can be viewed on different levels to understand the whole process in a graphical view. These workflows, backed by various graphical analytical charts, provide detailed information on bottlenecks and the steps that require to improve in the process by automating them. Business analysts and technical experts can coordinate and come up with the best approach to optimize and automate such steps by analyzing the data provided by UiPath Process Mining.

Graphical representation of the process — (Source)

For anyone new to understanding the process that’s in the pipeline to automate, UiPath Process Mining provides a good start as it is capable of answering many questions like:

  • What happens in a specific business process?
  • What are the root causes of the inefficiencies?
  • What are the deviations from the intended process?
  • What is the cost of the process?
  • How to optimize the process?

On the other hand, if the organization has a requirement to automate a specific process, the use of process mining could help them understand the process better. Using this tool dramatically increases the efficiency of this time-consuming process to analyze the process that needs to be automated. Proper detailed end-to-end understanding of the process allows us to determine the required input data and output data quickly. It is the understanding that require critically for the automation.

Business analysts and Solution architects spend countless hours interviewing and analyzing the data provided by the users who perform specific processes. But later, figuring out the provided information does not cater to the full requirement, and the developers are unable to continue with the development results in extra effort on capturing the missing elements at different phases of the journey. Occasionally, finding a critical component that remained unable to discover previously could result in redesigning a significant portion of the automation solution. Such incidents could significantly impact the project timelines and the cost involved. Hence, the recommended method is always to capture the detailed step by step requirement and cross-verify it with the user who performs the process, and with the management before getting into standardizing or automating the identified process.

UiPath Automation Hub is a great tool to share and showcase the automation ideas. Process discovery is all about looking for the best suitable process to automate. UiPath Automation Hub allows the users to share their automation ideas while others in the organization can vote for the opinions published. While the users can view and suggest their thoughts on the automation ideas, the UiPath Automation Hub also calculates the automation potential of the submitted idea by asking several key questions from the person who provides it. The scores and rating supplied by UiPath Automation Hub gives a great insight into prioritizing the automation requirements based on the complexity, automation potential, and the return on investment. It is a fantastic way of capturing the automation requirements across the organization by getting the employees in different divisions to collaborate to get the mundane tasks automated. Through Automation Hub, the business users can easily record the steps performed to complete the process using the UiPath Task Capture. UiPath Automation Hub enables the creation of an RPA Center of Excellence (CoE) within the company, which focuses on building a long term automation pipeline. Once the automation pipeline is ready, the respective employees in the CoE could pick the high priority processes from the queue and dig in for more details by getting in touch with the individual users who posted the requirement.

UiPath Task Capture allows the user to record each step involved in the process seamlessly. Once the recording is complete, UiPath Task Capture allows the user to automatically generate the Process Definition Document (PDD) to support the business analysts and development teams.

UiPath Task Capture — Recorded workflow

UiPath Task Capture dramatically reduces the time and effort business analysts put in to record and document the as-is process. Task capture is capable of capturing all the steps the user performs and generates a detailed document that consists of screenshots, diagrams, and many other information along with the time spent on each screen to calculate the total time spent to complete the process when performed manually. Not only that, but it also allows the developers to extract specific workflow snippets from the captured process to open it through UiPath Studio directly. The document and the workflow snippets provided by UiPath Task Capture acts as an excellent foundation to start building the automated workflow without much effort.

These applications make the process discovery quite easier, beneficial, and efficient than performing it manually. The next most crucial step for the business analysts are to get the requirement finalized by presenting it to the business users and their respective management.

Requirement Analysis and Finalizing

The finalized requirement has to be analyzed by the business analysts and the respective technical leads like RPA solution architects to look for opportunities to standardize the existing business process before applying automation.

How to standardize a business process?

Process Standardization with RPA Document

Building the To-be Process

RPA solution architects and RPA team leads should work together with the business analysts to understand the requirement and the as-is process map, which business analysts create after understanding the full specification. The solution should be designed based on the new and optimized process focusing on the most efficient delivery of the process. The suggested business process optimizations and standardizations must get communicated to the business users and the respective management to get the confirmation on the process changes. The Process Definition Document (PDD) should consist of all these changes under the proposed solution.

Building the To-Be Process Map

The to-be process design should consist of the most granular level steps. In other words, the diagram should include all the actions, decisions, and sub workflows needed to complete the process. The PDD may contain multiple to-be process diagrams for one business process because the process may break down into various sub-processes that run parallel to other automation tasks or sequentially. The detailed, granular level diagram should identify the workflows needed to handle each task. The overall effort is estimated based on the complexity of the steps (technical and business logic) involved in each designated workflow task.

To conclude, process discovery is a crucial part of any RPA project. Process discovery includes requirement gathering, analysis, and documentation as few primary steps. The inability to perform proper process discovery could lead to increased costs and ending up with ineffective automation solutions. The most crucial point is that the process discovery scope is much broader than a software solution requirement discovery. Hence, to capture the proper requirements effectively and efficiently, the use of appropriate techniques is mandatory.



Lahiru Fernando

UiPath MVP/ Executive Lead — Robotic Process Automation/ UiPath Community Leader for SAARC Region/ UiPath Community Moderator