Working with UiPath Process Mining: How to Get Started

man looking at sticky notes on wall

Process mining enables organizations to reveal invisible processes by analyzing event logs in IT systems to automatically map the ‘as-is’ process (as opposed to the ‘to-be’ process), and measure important business process metrics related to time and quality (such as throughput-time, process cost, and efficiency).

If you’re new to the idea of process mining, we suggest you start by reading this article to better understand the basics of process mining and the value it provides to organizations.

Process mining is a powerful tool in the right hands - but like all tools, there is a right place and procedure to use it effectively. To help you understand these nuances, in this post we’ll discuss how to qualify and target a business process for UiPath Process Mining. I will also provide you with an understanding of where and how it makes sense to use UiPath Process Mining.

Step 1: Understand the goals you want to achieve

First, look at your organization’s department and get an understanding of the ‘pain’ that your business is feeling and/or what problems you are facing. This pain can be in relation to time, cost, and/or quality that a business process is causing. 

Understanding the pain will help you better qualify the process that needs to be analyzed. The next steps will cover how you will use UiPath Process Mining (available via UiPath Platform) to solve the problem you identified. 

Step 2: Qualify a batch of processes

Once you get an understanding of the potential processes that are causing problems in your department, the next step is to qualify the batch of business processes and scope them based on their dependencies. There are several ways to understand if your business process qualifies as a good candidate for process mining:

System type: Describing if the business process logs are stored in modern applications or legacy applications. This will directly affect the extract, transform, and load data (ETL) methodology, as you may have to use customized Python or R scripts to scrape data in legacy systems or simply make a SQL query to a database.

Data volume: This is a key identifier for complete data mining. UiPath Process Mining works to find patterns in the logs to automatically map and creates the process. If there is not enough data, then the process graph that the software builds will represent only part of the process.

System dependencies: It is important to understand the IT system landscape of your organization and how many applications your business process touches. If you are dealing with 10 applications, then you would have to connect to all the systems to be able to build the most accurate and complete process model. The more applications your business process uses, the more effort there is to extract, transform, and load the data into UiPath Process Mining.

Analog or digital: Similarly, the business process you are going to analyze should have a digital footprint. If you are unable to track the process logs, then you may be better off analyzing the process traditionally or with UiPath Task Mining.

Number of subject matter experts (SMEs) and time: Make sure that the business process is executed by many SMEs over a long period of time. If there are only a few people involved in the process, you may be better off just using business analysts to capture the process. Similarly, the process must be done over a significant period of time to accumulate the action repetitions and process variations. 

Step 3: Extract, load, and transform (again and again)

Once you decide which business process you are going to analyze, the next step is to go through the ETL process. Here it is very important to understand the underlying table names behind applications like SAP S/4HANA or Oracle. The better your understanding of the table names and where data is stored, the faster you will be able to complete this step. The event logs reveal the process graph. But to measure goals related to time and money you will need to extract key performance indicators (KPIs) that may be stored in another system. 

The attributes of the process are essential to measure the impact of the process. For that reason, it is highly important to understand which application the business process touches. The goal here is to track the CASE ID across multiple systems and understand how users interact with them, therefore building the system story  (a short summary of the whole project). 

Then, you will find yourself extracting revenue data from SAP and fully loaded full-time equivalent (FTE) costs from Oracle. Meaning that you may iteratively go back to the applications and look for the data field you really need to find the solution to the initial problem your business defined. UiPath Process Mining offers “Look Up” functionalities to map the data onto each other and based on the Key Identifier, also called CASE ID, you will be able to track the impact of any individual transaction item. 

Step 4: Understand the happy path vs. most frequent path

Process mining will generate the process graph based on the most frequent path with core activities (note that the displayed graph is not always the “happy path”). While conducting your conformance checking, it is very important to compare the two process variants, and this will help you identify areas for improvement. If the current process does not conform to the happy-path process graph, then you know that your process is not being executed as it is documented or planned. 

Step 5: Defining the solution and business case

The goal of a process mining project is to find a solution to the initial business problem selected at Step 1. At this point, you will use the UiPath Process Explorer dashboard and start digging into the different process variants and deviations. Doing so will help you understand how you can achieve the initially-set goal based on your new and enhanced understanding of the process. You will be trying to find ways to achieve your goal. 

A goal can be any goal that is relevant to your business, like “decrease process throughput time,” “decrease invoice-delays,” “increase efficiency in delivering orders,” or “decreasing process cost related to time.” To accomplish these, you would need to conduct an analysis to identify the specific activities that are bottlenecks. This will allow you to present a targeted business case backed with measured as-is process metrics. 


Finally, once you identify these activities (above), the next step is to decide on the solution or the remedy to improve the business process. The solution may entail implementing a robotic process automation (RPA) solution, or optimizing and redesigning your process. Whatever the choice, UiPath Process Mining excels in creating a holistic view of the business process and targeting the areas of improvement with a “top-to-bottom” approach. 

The step-by-step guide above demonstrates how to quickly and effectively implement process mining. The ability of the tool is quite extensive and can be used to continuously improve processes. After you implement your solution, you will be able to go back and measure the delta, the impact of your changes (or lack thereof), and validate the accomplishment of your business case.

To learn more about UiPath Process Mining, check out the UiPath Academy course.

Avatar Placeholder Big
Alp Uguray

Head of UiPath , Symphony Ventures