FortressIQ just comes out and says it: To really understand business processes, feed your staff’s screen activity to an AI

In a sign that interest in process mining is heating up, vendor FortressIQ is launching an analytics platform with a novel approach to understanding how users really work – it “videos” their on-screen activity for later analysis.

According to the San Francisco-based biz, its Process Intelligence platform will allow organisations to be better prepared for business transformation, the rollout of new applications, and digital projects by helping customers understand how people actually do their jobs, as opposed to how the business thinks they work.

The goal of process mining itself is not new. German vendor Celonis has already marked out the territory and raised approximately $290m in a funding round in November 2019, when it was valued at $2.5bn.

Celonis works by recording a users’ application logs, and by applying machine learning to data across a number of applications, purports to figure out how processes work in real life. FortressIQ, which raised $30m in May 2020, uses a different approach – recording all the user’s screen activity and using AI and computer vision to try to understand all their behaviour.

Pankaj Chowdhry, CEO at FortressIQ, told The Register that the company had built was a “virtual process analyst”, a software agent which taps into a user’s video card on the desktop or laptop. It streams a low-bandwidth version of what is occuring on the screen to provide the raw data for the machine-learning models.

“We built machine learning and computer vision AI that will, in essence, watch that movie, and convert it into a structured activity,” he said.

In an effort to assure those forgiven for being a little freaked out by the recording of users’ every on-screen move, the company said it anonymises the data it analyses to show which processes are better than others, rather than which user is better. Similarly, it said it guarantees the privacy of on-screen data.

Nonetheless, users should be aware of potential kickbacks when deploying the technology, said Tom Seal, senior research director with IDC.

“Businesses will be somewhat wary about provoking that negative reaction, particularly with the remote working that’s been triggered by COVID,” he said.

At the same time, remote working may be where the approach to process mining can show its worth, helping to understand how people adapt their working patterns in the current conditions.

FortressIQ may have an advantage over rivals in that it captures all data from the users’ screen, rather than the applications the organisation thinks should be involved in a process, said Seal. “It’s seeing activity that the application logs won’t pick up, so there is an advantage there.”

Of course, there is still the possibility that users get around prescribed processes using Post-It notes, whiteboards and phone apps, which nobody should put beyond them.

Celonis and FortressIQ come from very different places. The German firm has a background in engineering and manufacturing, with an early use case at Siemens led by Lars Reinkemeyer who has since joined the software vendor as veep for customer transformation. He literally wrote the book on process mining while at the University of California, Santa Barbara. FortressIQ, on the other hand, was founded by Chowdhry who worked as AI leader at global business process outsourcer Genpact before going it alone.

And it’s not just these two players. Software giant SAP has bought Signavio, a specialist in business process analysis and management, in a deal said to be worth $1.2bn to help understand users’ processes as it readies them for the cloud and application upgrades. ®

Source: FortressIQ just comes out and says it: To really understand business processes, feed your staff’s screen activity to an AI • The Register

Robin Edgar

Organisational Structures | Technology and Science | Military, IT and Lifestyle consultancy | Social, Broadcast & Cross Media | Flying aircraft

 robin@edgarbv.com  https://www.edgarbv.com