Process Mining: Deriving Actionable Insights

Microsoft's Power Platform has now become a full hyperautomation stack thanks to the recent general availability announcement of Process Mining!

The Process Mining capability works with your systems of record to visualize processes happening in your organization, ultimately allowing you to gain insights into these processes and identify areas for improvement.

Process Mining has several components to help us better understand a process, such as:

• Process maps for visualizing the process
• Charts summarizing key statistics related to the process
• A root cause analysis component for finding hidden connections in the process data
• A summary of variants to understand the different paths of a process
• A process comparison component to identify similarities and differences between processes
• And more!

In this article, we are going to walk through an example of how someone might logically use Process Mining to gain actionable insights into their process.

Using Process Mining on an accounts payable refunding process

You’re an account manager for Contoso and want to better understand your current processes. Specifically, you want to look into your account payable refunding process, which tends to be lengthy and complex.

Process event logs

You’ve pulled the event logs for your accounts payable refunding process. This data will be the input for creating a new process mining instance.

Process summary

After uploading the data into the Process Mining tool, you can see a high-level report of this process.

• Activity: A unique event/action within a process
• Case: An instance of a process, where at least one activity occurs
• Variant: A sequence of activities that make up a unique path for the process

Process map

The process map helps us visualize the steps in a process.

A few things this map is telling us:

• Exhibit A: Numbers within the activity circles - number of cases that include that activity (i.e., there are 68 cases that include the "Refund Customer" activity)
• Exhibit B: Blue ripples around the activity circles - relative count of cases that include that activity (i.e., "Refund with Standard Voucher" occurs relatively often in cases)
• Exhibit C: Numbers next to arrows - number of cases that move from one activity to the next (i.e., of the 65 cases where "Confirm Payment Received" occurred, only 9 of them went on to the "Refund with Special Voucher" step)

Total cases

You notice that Invoice Entry and Re-issuing the invoice have the highest number of cases, because they are the start/end points of the process.

You also notice that Complete the Customer Memo (CCM) has a high case count too, at 113.

Total duration

You change the map’s metric to total duration.

You are surprised to find that Complete the Customer Memo (CCM) has a relatively high total duration - almost 3 weeks.

You wonder: the Complete Customer Memo activity occurs often and appears to be time-consuming...is it worth exploring this activity further?

Total invoice value

You change the map’s metric to total invoice value.

Of the $182k that flow through this process, about half ($90k) is involved in the Complete the Customer Memo (CCM) activity.

Given how much of the total invoice value flows through this activity, you want to take a closer look and understand the activity's impact on the process.

Filtering process views

You create two views of this process:

With the CCM activity

Without the CCM activity

Now, you want to compare these two views.

Process comparison

Activities with higher (positive) numbers: Activities that tend to only occur with Customer Memo activity

Activities with lower (negative) numbers: Activities that tend to only occur without Customer Memo activity

It has become quite obvious that cases with the CCM activity are much more complex.

You ask yourself: could CCM be causing this complexity?

Drilling into CCM

You select the CCM activity and see that nearly 43% of cases involving the CCM activity need to be redone.

You want to understand why cases with CCM end up requiring so much rework.

It looks like Addison and Jackie are the employees who perform this task the most.

You set up a meeting with them to discuss the customer memo activity.

Process Mining Copilot

Before your meeting with Addison and Jackie, you turn to Process Mining Copilot to confirm your findings.

Copilot provides the top insights, mentioning the CCM activity as repetitive and lengthy.

You feel confident that the CCM activity is a bottleneck…now you need to figure out why.

Identifying the bottleneck cause

After meeting with Addison and Jackie, you find out that the customer memo step is being done manually by copying customer information from a spreadsheet into the customer memo document.

They also mention that manually copying information over often leads to data or formatting errors, causing the approver to reject and send the invoice back for changes.

You realize that this activity needs a solution that will standardize data entry, speed up customer memo creation, and streamline the approval process.

Putting the actionable insights to work

To standardize and speed up the customer memo step, you set up a Power Automate workflow that automates Addison and Jackie’s manual tasks while streamlining the approval process.

A few weeks later, you discover that this automation has reduced the total process time by 15.31 days – over two weeks’ time saved!

This is a basic example of how we can use Process Mining to understand our process, extract key insights, and take action on process improvement.

Try out Process Mining today!