Process Mining Use Cases

Process mining is helpful in many scenarios, whether you want to visualize your process for the first time, uncover more detailed insights into compliance or efficiency, and much more. As long as your processes have data to back them up, there will be something new to learn and efficiencies to find.

This page shares some common use cases that process mining can address.

• If you already know what you want to do, see when to use process mining.
• If you need more inspiration, see use cases.

When to use Process Mining

I want to…	How Appian can help
Transform existing process data into a format that Process Mining can analyze.	Use Mining Prep to transform data into event logs and upload them into Process Mining to start your analysis.
Visualize a process to gain new insights into how a process actually executes.	Upload an event log and Process Mining automatically generates a discovered model. Here you can adjust a variety of visual settings and filters to hone your analysis.
Track, analyze, and share performance trend summaries for multiple processes.	Use process scorecards to report on performance trends and health for your mined processes. Here Appian scores your processes based on custom performance objectives that you configure.
Visualize, monitor, and share process metrics as charts and graphs.	Create dashboards to display process metrics as key performance indicators (KPIs), histograms, bar charts, and more.
Check how a process conforms against a target model.	Add a target model and Process Mining finds the differences between the discovered and target models. View precise conformance statistics on the Statistics page.
Identify issues in a process that negatively impact its performance.	Use root cause analysis from the Insights page to uncover deviations such as skipped events, events with switched orders, and more.

Use cases

Optimize manufacturing process

Your car manufacturing plant wants to identify optimization opportunities.

• As a business analyst, you want to identify optimization opportunities so that you can implement data-driven recommendations to improve the efficiency of the manufacturing process.
• As a transformation developer, you want to quickly transform process data from your order tracking system into event logs so that your business analyst can uncover optimization opportunities from the data with Process Mining.
• As an executive overseeing the manufacturing plant, you want to review high-level results from the process optimization efforts so that you can ensure your processes are running at maximum efficiency.

Due to the complexity of the process and the sheer volume of production, it is challenging to identify process bottlenecks and inefficient deviations. Fortunately, your order tracking system records data about each order as it passes through the assembly line from order creation to delivery completed. This data is your data set, which is full of opportunities to uncover, but you must first transform it into an event log so it is suitable for Process Mining analysis.

Mining Prep is a powerful, low-code tool that you can use as a transformation developer to transform your data set into an event log. In other words, you need to make sure that you identify all the required elements within the data set such as the case ID, time stamps, and other attributes. For example, your plant uses each vehicle's VIN to track it through your manufacturing process. In this case, the VIN can serve as the case ID. The case ID maps events, such as Order Created, Chassis Completed, and Body Painted to a single case or single occurrence of a process. Once you're ready, you can transform the data set into an event log and load it into Process Mining.

Now that your event log is in Process Mining, as a business analyst you can view a visualization of your process in the form of a discovered model. By default, the discovered model will show you the most common variants or paths your process takes, including how many cases went through each activity, how long each activity took, and the wait time between activities.

Within the discovered model, you notice there is a longer than expected wait time before your Quality Control event, which is causing a bottleneck in your process. The following screenshot shows how connectors between activities are red when there are high average activity durations.

Using the metrics on your dashboard and statistics pages, you identify that the case duration is trending upwards over time, but quality control personnel are working at maximum capacity.

To further confirm your assumptions, you run a root cause analysis on the high case durations and find that they are highly correlated with event attributes that relate to quality control workers.

Now you know what is probably the source of your process bottlenecks, and you determine that you need to hire more resources to alleviate the bottleneck and reduce cycle times.

After you've implemented changes to your process, you can repeat this procedure again and again to verify the results of your optimization and continue uncovering more opportunities for improvement. To convey the results of your process improvements, you can create a process scorecard and share it with the business executive so they can track the performance of this process over time.

Internal audit of P2P process with root cause analysis

Your Procurement department wants to reduce maverick buying and increase compliance for a procure-to-pay (P2P) process.

• As an auditor for your Procurement department, you want to identify and resolve instances of maverick buying across the organization so that you reduce costs, increase quality of purchased goods, and improve contract compliance.
• As an Appian Developer, you want to implement process improvements that the auditor identified so that you reduce maverick buying.

As an auditor within the Procurement department, you already have a target or ideal process in a BPMN format. After you upload the associated event log into Process Mining, you can import the BPMN target model. This provides a reference point for Process Mining to compare your actual and target process.

Once you've uploaded the event log and target model, Process Mining automatically visualizes your actual process as a discovered model and your target model as a BPMN model. A good first step is to look at the target model which overlays deviations in your process.

Here, the target model indicates that the Purchase Order: Approved activity is skipped in several cases. You know that the activity is skipped because the red dotted line shows an alternate path, and the Activity Skipped icon displays on the activity. This is a deviation in the process that you want to resolve, but you still aren't sure how to start fixing this issue. This is where automated root cause analysis can help. Root cause analysis uses machine learning to uncover why a deviation occurred.

The Deviations page displays all process deviations between the actual and target process, such as the skipped Purchase Order: Approved activity that leads to maverick buying. After you perform root cause analysis, Process Mining identifies a pattern that the majority of these cases came from a specific department. This indicates to you that this department needs more communication about the correct process, and they need an easier way to align with that process.

Thankfully, your Appian developer can now leverage the full suite of Appian functionality to streamline and enhance this P2P process with business rules, process models, or other automation capabilities to remove the deviations. As you continuously analyze your process data in Process Mining, you can see at a glance how your changes are improving the process or where more opportunities still exist. As you discover more about your process, you can create personalized dashboards that track key metrics about your process, which further help you pinpoint areas for optimization.

Feedback