In this article, we’ll explore what process mining is, what it can (and can’t) do for businesses looking to optimize their processes, and how process intelligence offers a more effective approach.
What is process mining
Process mining uses actual data from information systems to create a model that accurately reflects how a process executes.
Applications such as CRM and ERP systems, as well as other systems of record, automatically create event logs that record every action taken. The data in these logs can be collected, or “mined,” to create an audit trail of the processes the applications are involved in. This works even when multiple applications are used in a single process. Process mining technology follows these audit trails to build a process model showing the details of the end-to-end process, as well as variations. Business users can analyze these models to find out if the processes are functioning as they should and, if not, investigate the root causes of deviations from the optimal path.
Businesses live by their processes—the prescribed sets of actions their employees take to get things done. When processes run well, the business runs well. When processes run poorly, the business risks a host of hazards, from loss of revenue to customer dissatisfaction to compliance violations. Most businesses have a general idea of how their processes should run, but lack insight into the day-to-day details of execution. Without this insight, how can they make improvements that yield real results?
Process mining offers one solution, and for many years it served businesses well. However, in today’s increasingly complex environment and amid growing pressure to do more—faster and at lower costs—organizations need intelligent automation solutions.
How process mining works
Before process mining, the only way for businesses to analyze the performance of their processes was through interviews with business users and manual data reviews—a slow, tedious undertaking with a high margin of error. Process mining allows organizations to leverage automation to paint accurate pictures of real-world process performance—faster, easier, and more accurately than manual approaches.
Where process mining falls short
Process mining offers enormous advantages over manual approaches to process analysis, but it has its limitations. For example:
- Traditional process mining identifies process-related issues, but stops short of providing granular answers concerning the root causes of those issues.
- Process mining works well in simpler scenarios, but lacks the sophistication to evaluate complex processes with a large number of valid variations.
- Process mining can only analyze past performance, lacking the ability to monitor processes on an ongoing basis and to alert users to deviations.
- Some traditional process mining tools may be limited in the types of data sources they can connect to, which can limit the value they can provide.
Process intelligence bridges the gap
A new generation of process analytics solutions goes beyond traditional process mining. Process intelligence combines business intelligence-like metrics with a set of process-specific analytics to offer detailed insights into complex processes from end to end. Unlike traditional process mining, process intelligence enables businesses to view their processes in real time and analyze patterns that lead to bottlenecks or disruptions.
Here are the top five advantages of process intelligence versus traditional process mining:
1. Timeline-based analysis
Process mining uses the schema method of process analysis, which involves converting process data into a flowchart (schema) and then analyzing the flow of all iterations through that schema. The shortcoming of this approach is that few business processes fit into a well organized flowchart. By the time all valid variations to a process are considered, the schema often becomes a tangled mess with limited usefulness.
By contrast, process intelligence uses a timeline approach, which creates an unfiltered, unedited visual history of every process end-to-end, even when some steps are performed using multiple systems. These timelines are then analyzed so that they can be compared, filtered, searched, aggregated, etc., similarly to how a business intelligence (BI) application analyzes records in a table. And Primary Path view works equally well on all types of processes, finding deviations from the common flow, the reasons for such deviations, and revealing them from different perspectives. As compared to basic process mining, that only works well on processes with little variability in terms of the sequence of steps, process intelligence ensures that digital transformation initiatives deliver predictable results and do not cause unintended consequences.
2. Plan for success with process simulation
Now you can utilize machine learning to improve and optimize your processes. No other solution offers the same amount of process insight and analytics on one platform. With predictive and prescriptive analytics in the new Process Simulation tool, you can simulate potential changes in a process and evaluate the impact that the changes will have on the entire business process, before you make them.
3. Continuous improvement
Traditional process mining is focused on looking at historical data. While this approach can offer valuable insights into what worked well and what didn't, it falls short of offering solutions for present and future iterations.
Process intelligence monitors processes with new data coming in real and near-real time, “watching” every iteration and alerting process owners for deviations that could cause delays or other problems. By enabling continuous improvement, process intelligence continues to deliver ROI as businesses capitalize on new opportunities to make processes work faster and smarter.
4. Reduces compliance risks
When businesses run traditional process mining applications, users can review the output to identify present and past deviations that could lead to compliance issues. This approach relies on the expertise of the users reviewing the data.
Process intelligence enables users to define process rules that align with the organization’s compliance requirements and to instruct the system to watch for violations. When one or more of those rules is broken, the system alerts users right away, enabling them to take immediate action to rectify the deviation and to ensure that it will not happen again. Process intelligence alerting rules can also be defined to call a service when an alert is triggered, to automatically deal with the problem. This capability can mean the difference between discovering an issue just in time, before it affects a business’ compliance status, and finding it when it is too late to be fixed and has already caused problems elsewhere in the workflow—or worse, learning about it after a violation has been reported.
5. RPA enhancements
According to Ernst & Young, between 30 and 50 percent of initial robotic process automation (RPA) projects fail due to lack of quantifiable process data. As businesses deploy RPA for more intricate processes in more complex environments, the pressure to deliver positive ROI has increased dramatically, and traditional process mining can offer only limited support in yielding the returns that businesses are seeking.
Fortunately, process intelligence can be just as valuable to digital workers (RPA "bots") as it is to human employees. Today’s process intelligence solutions can include process mining and task mining. As with process mining, task mining is looking for the significant events in a process. Task mining adds the ability to record a user’s manual actions on their computer to capture manual process steps to be used alongside the steps gleaned from system of record log files. By applying process intelligence to manual as well as automated processes, businesses uncover new opportunities to improve RPA results:
- Evaluate manual tasks as to suitability for automation
- Capture the steps in a manual task and use this as a template to create the required bot
- Spot previously hidden redundancies
- Identify optimizations that can free up digital worker cycles, improving digital workforce productivity
- Discover and remedy inefficient hand-offs between digital and human workers
- Deliver quantifiable data on the financial impact of digital workers by process
- Compare human and digital labor in terms of cost, accuracy, efficiency, and duration
Why process intelligence is the future of process improvement
For many years, process mining applications served process owners well, saving countless manual hours and helping businesses discover opportunities for improvement. Process intelligence provides a new approach to process improvement that improves upon process mining. Process intelligence works with all processes, simple and highly variable, manual and automated. Process intelligence will monitor every process instance as each new step occurs, alerting or even taking automated action whenever a process behavior of interest is seen.
Process intelligence supports RPA initiatives by identifying good automation candidates and then monitoring and reporting on the process that bots participate in. Process improvement can now reach a new level in delivering on its promises of greater productivity, reduced risk of costly compliance violations, and the streamlined efficiency that can create happier customers, happier employees, and a greater competitive edge.
Ready to leverage the data you already have for business process improvement? Click the button below to get started.
Editorial note: This blog post was originally published on July 21, 2021. Updated on March 30, 2023, to reflect the new capabilities of ABBYY Timeline 6.0 including new UI, deeper analytics, and process simulation, further advancing ABBYY's process intelligence capabilities.