How Process Mining Uncovers Workflow Bottlenecks

To improve efficiency, agencies must understand their processes, not just the technology that performs them, an expert says.

The way to do that is with process mining, which, according to IBM, is “a technique that applies specialized data mining algorithms to system processes in order to better understand workflows, uncover bottlenecks and gain actionable insights into process management.”

The technology was developed because processes have become more complex, said Mike Daniels, senior vice president of Global Public Sector at Celonis, which makes process mining and execution management software. 

“You start losing visibility of how things actually flow through your system,” Daniels said. “You have an idea of how they’re supposed to flow through your system, you have a stated ‘this is what it should be,’ but how it actually flows through the system gets complicated to see,” he said. Processes cut across many technology stacks and departments, so there’s “very little visibility into the end-to-end process,” he added.

Process mining is particularly important to the public sector, Daniels said, because agencies have invested in modernization and system transformation, but systems are often siloed, which clouds visibility into processes. Potential use cases in government include service or benefits applications, building maintenance and financial management.

“The value proposition is really all about ‘Hey, you’ve probably spent a fair amount of time as an agency investing in some sort of digital agility. You’ve probably done a number of system transformations over the years,” he said. Process mining is about having a tool that lets you see how to optimize on those investments by understanding how they impact processes, he added. 

Additionally, benefits of process mining include cost savings, operational efficiencies, workforce productivity and better customer experience, according to the Everest Group, a research firm – all of which top public-sector priorities lists.

A Dutch computer scientist coined the term “process mining” in 1999, but the technology is gaining traction now. Process mining is part of hyperautomation, which lets organizations identify, vet and automate processes, and Gartner predicts that the market for the software enabling this will reach nearly $860 billion by 2025. Polaris Market Research estimates the global process mining software market to reach $11 billion by 2030, up from $374 million in 2021.

According to a Forrester report commissioned by Celonis and released in January, 61% of decision-makers will use, or are evaluating, process mining in the next 12 months, ranking it the top technology they plan to use to measure or improve their business processes. 

What’s more, big tech firms are investing in it. IBM helped a state government implement a process mining solution to enable two SAP products to exchange data in real time. The result was a decrease in process lead time from five days to 15 hours.

At the end of March, Microsoft acquired Minit, a Slovakia-based process mining tech vendor, signaling “Microsoft’s commitment to help organizations quickly discover and optimize their business processes by bringing data and execution together to unlock powerful insights,” according to a press release.

Also in March, Celonis announced the acquisition of Process Analytics Factory, a provider of process mining insights for Microsoft Power BI. The deal will enable Microsoft Power Platform users to use Celonis’ Execution Management System for process mining. Earlier that month, a company blog post targeted the public-sector market as ripe for reaping process mining’s benefits.

Celonis technology creates a digital twin of an agency’s processes that shows whether processes are meeting key performance indicators. The data is collected in real time to provide continuous visibility into processes.

As a result, data and patterns begin to emerge such as “what percentage of the time does it actually go through what we would call the happy path – what you think it should follow – vs. what percentage of the time does it go off?” Daniels said of a process. “Process mining is aimed at making that invisible visible.”

Agencies can then manipulate the digital twin of the processes to determine how best to tweak the processes, rather than making changes to the processes themselves without knowing how they’ll affect them.

“It creates this level playing field where you as the agency see the world exactly the same way I do because it’s digital truth,” Daniels said. “It’s not subjective. It’s all driven by the data that comes up.”

Stephanie Kanowitz is a freelance writer based in northern Virginia.