Banks and financial institutions want to automate back-end processes but very often the conversation can get confusing right at the start. Realizing that automation adds efficiency is the first step, but figuring out what to automate first — and how — can be tricky.

On a regular day at any institution, humans and software applications work in conjunction to get things done. While most software logs everything it does by default, asking a human to record each step delivers mixed results.

Therefore, when financial institutions exploring what processes they should automate, two approaches can be helpful: process mining and process discovery aka task mining.

Process mining involves looking through software event logs to understand how business information is processed. Process discovery, alternatively, looks at a user’s interactions with software by recording and mining the logs at the user-interface level.

Process mining

Process mining has been around for about 15 years, Craig Le Clair, principal analyst at research firm Forrester, told Bank Automation News, but until recently, the limited return on investment from doing a comprehensive audit of system logs limited the process to “a bit of a science project,” he said.

But with the rapidly increasing use of systems based on robotic process automation (RPA) and artificial intelligence (AI), the ability to gain a comprehensive view of business processes has become quite relevant.

With process mining, organizations are “able to spot shadow processes, bottlenecks and avoidable delays occurring due to manual inefficiencies,” noted Tushar Shrivastava, an analyst at research firm Gartner, in a research note shared with BAN.

Process discovery

Process discovery, on the other hand, allows a business to learn which repetitive tasks employees might be performing at their workstations. The process looks at “human-to-machine interactions,” Le Clair said, adding that the log of interactions is then moved to machine learning engines, which can help detect patterns and identify redundancies.

Process discovery has also garnered attention from venture capital firms like Citi Ventures, which has invested $14 million into AI observation-based software provider  Skan. “We’re excited about companies in the process discovery space that allow you to automate the discovery element from the perspective of the person running the process at the computer terminal,” Matt Carbonara, managing director at Citi Ventures, previously told BAN.

The data collected by process mining and discovery not only helps institutions understand what they can automate, but also “informs the building of the bot,” Le Clair said, since the process information can be added to software’s wireframe and help speed up the automation’s implementation.

Best practices

Basic process mining best practices, according to Gartner research are:

• Extraction of relevant data, such as unique case IDs, time stamps and activity, from IT systems;
• Input of data into a process mining engine; and
• Generation of a process flow chart.

While process discovery can deliver insights on how standardized tasks are completed by individuals, it can also capture “noise,” at times when a “user is interrupted or gets distracted during the execution of a task, leading to performing activities that are not relevant to the task in question,” according to Srivastava’s Gartner note. “A process mining initiative doesn’t need all the data about all the tasks in a process at once,” he said, adding that identifying easily available data about processes can also deliver key insights.

With an increased focus on how and what to automate, RPA vendors such as UiPath and Automation Anywhere are also offering business process mining tools. Other notable vendors in the process mining space include Celonis, Cognitive Technology, Fluxicon, Icaro Tech and Kofax. In fact, Celonis, which has a functional partnership with Automation Anywhere, recently announced a blockbuster billion dollar funding round which parked the company’s post-infusion valuation at $11 billion.

That massive funding round, along with increased interest in businesses that provide process discovery, highlights the renewed attention automation techniques have brought to raising efficiency and understanding workflows within banks and financial institutions. In other words, process discovery is “RPA raising the tide for all boats,” Le Clair said.