Process mining software pulls back curtains on how processes work

Task mining fills the gaps in processes

One of the leaders in developing a process mining application is Celonis, a Munich-based firm.

The Celonis core process mining software integrates into ERP and other enterprise systems to capture key strokes and screen activities as a process happens. The data is analyzed to map out the process from start to finish. However, process mining only captures what's happening within the enterprise systems, and a process usually involves activities or steps that occur outside of these systems, said Alexander Rinke, Celonis CEO.

Alexander Rinke

Celonis recently added a Task Mining Engine to its AI driven process mining engine, which fills in the gaps of process mining and provides an even more complete picture of a process by including human factors, Rinke said.

"So far, [Celonis process mining] focused on the business data in the company -- the orders, invoices, shipments which start in the business systems like CRM or ERP," he said. "However, we missed anything that was happening outside of the systems, such as somebody leaving the ERP to open up an Excel file or to send an email. The new Task Mining Engine adds those pieces to the puzzle, and it can identify everything that happens on people's workstations outside of the business systems."

For example, the Task Mining Engine can help identify how much time customer-facing employees spend on enterprise systems and how much on customer contact, or how and when procurement workers go outside of a procurement system to communicate with vendors.

Uncovering hidden parts of processes

Task mining is valuable for organizations that want to introduce RPA and therefore need to know how tasks fit into larger processes, said analyst Maureen Fleming, program vice president at IDC in Framingham, Mass.

"The thing about tasks is that they're fairly invisible because a lot of times a task might be something that you have to execute outside of an enterprise application like Salesforce or SAP, or it's a part of a custom workflow," Fleming said. "But what's actually done in that task is not completely understood, so the task mining comes in to help people say, 'Oh, this is how it works.' People have been looking for this for a while."

Celonis' task mining fits into an emerging trend of using AI more aggressively inside of process automation initiatives, she said.

"Celonis looks at process mining and task mining as one piece of what they offer for automation planning, and then they also tie that to predictive analytics to help understand what a process is about or whether a task could be automated using AI," Fleming said. "[Celonis has] a significant vision in how to build out these high level uses of AI in IT within processes."

Celonis' Task Mining Engine is "relatively unique" and is a disruptive approach in a significant trend, she said.

"This is an emerging category and there's a lot of activity around task mining," Fleming said. "It's the baseline for figuring out how to make more sense about how people are going to use RPA, but at the beginning, it's the data layer that's required to get more heavily involved with AI."

Process mining meets modeling

Another contender in the process mining software market is Signavio, a decade-old firm based in Berlin. Signavio's process mining applications originated from research on enterprise process improvement at the Hasso Plattner Institute, said Signavio founder and CEO Gero Decker. The Hasso Plattner Institute is a computer science research university founded by SAP co-founder Hasso Plattner in Potsdam, Germany.

Signavio has no direct investment from SAP, although Decker worked briefly at SAP Australia in process orchestration research before starting Signavio in 2009. SAP is one of Signavio's largest customers, Decker said.

Gero Decker

"The goal is to help organizations understand, improve and transform all of their business processes," Decker said. "It's about operational excellence, which in some cases is about compliance, in other cases it's about managing large-scale IT transformation projects like an SAP S/4HANA migration."

Signavio's process mining application can help organizations take a "rear-view mirror" approach to business processes and also monitor and analyze processes in real time, Decker said.

"You can take data from the last few months and drive improvement potential, for example, what are the five most important or impactful things that we can change and optimize," he said. "You can also use it for live monitoring where you see how you are performing today, what are the cases that maybe start showing strange behavior that you may need to escalate."

One differentiator for Signavio is embedded modeling capabilities that allow companies to discover what the potential impacts of changes to processes might be, Decker said.

"You can do simulation and forecasting as well, so if you say, 'Let's increase automation of the task or the process from 15% to 45%, what impact will that have in terms of bottlenecks or cycle times or capacity usage?'" he said. "This combining of modeling with process mining and analysis enables you to run a full circle on these process improvements."

Modeling and process mining are different but compatible, Fleming said.

"Process and task mining shows you how your process operates and modeling is about figuring out what are the things that you need to do to start working on how you want to improve the process," she said. "So in that sense, they work together really well for automation planning."