Process mining is a hot topic in many circles. Companies working on business process management, continuous improvement, procure to pay, and RPA are all interested in how process mining can help their organizations be more effective. The interest in process mining has led to incredible market growth over the last two years and a projected 200% annual growth moving forward.

The tremendous growth of the process mining market has led many software vendors to create offerings or re-message their current offerings toward process mining. Some of this messaging can be confusing, especially when vendors use the same words to describe very different capabilities. Fortunately, there are a common set of capabilities and definitions that have been standard in the industry for the last few years.

First, what is process mining? Gartner defines process mining as:

“Process mining is designed to discover, monitor and improve real processes (i.e., not assumed processes) by extracting knowledge from event logs readily available in today’s information systems. Process mining includes automated process discovery (i.e., extracting process models from an event log); conformance checking (i.e., monitoring deviations by comparing model and log); social network/organizational mining; automated construction of simulation models; model extension; model repair; case prediction; and history-based recommendations.”

To compare vendors, it is helpful to define common capabilities and determine which capabilities are needed for a particular use case. Here are ten (plus one) core capabilities to consider for your project.

1.      Automated Process Discovery – This is the primary capability of process mining applications. The ability to take a set of audit logs and automatically re-create the process in a compelling visual. The visual usually can isolate one or more process variants for analysis.

2.      Conformance Checking – Many use cases need the ability to compare the way processes are executed to approved execution models. Conformance checking allows you to compare an as-is transaction or variant to a pre-defined model or a process variant identified during process discovery.

3.      Data Preparation – To utilize the audit log data without impacting production applications, the data will need to be extracted from transactional systems. Once extracted, data must be transformed to cleanly associate attributes to specific cases and events. 

4.      Real-time Dashboards – Process mining is no longer solely focused on analyzing historical information. In today’s fast-paced world near real-time analysis is a necessity. Modern process mining systems should be able to ingest data in real-time and provide immediate analysis.

5.      Extended Applications/Actions – Monitoring process execution in real-time provides the opportunity to act on non-conformance as it happens. Process mining systems should have low-code/no-code action connectors to alert and/or remediate transactions that are veering off path.

6.      Process Model Enhancement – Discovered processes might need to be enhanced with additional data like which resource executed the transaction, what region/country/plant was part of the transaction, or which vendors were involved. Enhancing the process model with additional attributes allows for many of the advanced process mining techniques.

7.      Multi-Process Interaction - Process execution does not happen in silos. Many employees work on multiple processes simultaneously. The ability to identify the interactions and interdependencies between processes can be key to the success of some process mining use cases.

8.      Support for Customer Journeys – The past year has dramatically changed the way customers and companies interact in many scenarios. This is driving a huge increase in customer journey mining or visualizing the path customers take through websites using session logs.  Customer journey mining provides some unique challenges in capturing transaction ID’s and mapping customer journeys to internal processes.

9.      Predictive, Prescriptive Analysis – It is not enough just to know what happened in the past. Companies today need to know the root-cause of any issues, what is likely to happen next, and what they can do about it. That is the role of predictive and prescriptive analysis in next-generation process mining systems.

10.  Task Mining – Process mining uses transactional system audit logs to generate an overview of process execution. Task mining dives a little deeper into desktop-based events by recording what users do on their workstations. These recordings are especially useful in RPA use-cases to identify how the bot should be programmed.

Plus 1. Business Intelligence Integration – Most companies have invested a lot of time and money into advanced business intelligence applications like Tableau, Power BI, Qlik, or ThoughtSpot. Software has been purchased, users have been trained, data pipelines have been established, and culture has been developed. To increase adoption and reduce implementation time, process mining should not be a separate analytical application. Companies might evaluate whether there are process mining applications that integrate directly into their chosen business intelligence platform. It would be even more helpful if the existing BI platform was the process mining interface.

As you begin evaluating the use of process mining, it will be helpful to determine which of these capabilities are needed for your use case. You should then use that list to evaluate process mining vendors' offerings against your requirements. Defining the capabilities will help ensure everyone in the evaluation is speaking the same language.