KMWorld 2022 - November 7-10, Register Now !

To hyperautomate or not to hyperautomate?

Article Featured Image

A couple of years ago, the term “hyperautomation” was coined by Gartner. It was and remains a catchy term, and Gartner’s definition of what it means is pretty solid. To paraphrase, it uses AI and machine learning(ML) to automate processes and augment humans. But just like catchy news headlines, not everyone bothers to read and contemplate the underlying definition and rationale behind it.

Automate everything?

In our interactions at Deep Analysis with SIs (systems integrators), technol- ogy vendors, and users alike, hyperautomation is taken to mean “automate everything in sight.” With a recession looming and cost-cutting rising fast on the corporate agenda, many more will likely follow the path of trying to automate everything in sight. I say “trying to,” as most will fail miserably, and for a good reason: Many business activities cannot and should not be automated.

The name of this column is Ethical Innovation, and from the innovation standpoint, one would imagine that the world of enterprise automation is rife with innovation and new ideas. There is innovation, for example, the spotlight on task and process mining tools. In addition, 5 years ago, the growth of RPA tools certainly added new approaches to a traditional and somewhat staid technology sector. But as somebody who studies this world for a living, I can tell you with surety that there is more repositioning and rebranding than true innovation. Not that this is necessarily a bad thing. BPM (business process management) tools have been with us a very long time, yet still dominate the automation technology market. These tools work exceptionally well and will quite rightly be around for a long time to come.

From an ethical standpoint, automation has always trod a thin line, as there is no getting around the fact that automating with technology means less work for humans. It’s fair to say that enterprise automation is an ethical minefield, as constituency A wants to cut costs and improve efficiencies through automation. At the same time, the much larger constituency B needs a job to pay the bills. It’s such a minefield that I won’t wander blithely through it in this column other than to point out that it’s there.

The logic behind hyperautomation is clear: Automate everything that can be automated. The practicalities of that are far less clear. Consider, for example, the surge of interest over the past couple of years in task and process mining tools, which ostensibly enable you to automate more tasks and processes. This surge has, inadvertently perhaps, exposed many more challenges and, indeed, roadblocks, to automation than folks thought. In a real sense, task and process mining vendors are biting the hand that feeds them. Using these tools to understand how to automate a specific business process often exposes much more profound issues to be resolved. Furthermore, these same tools can suggest wrong paths to follow.

Pay attention to nuance

To dive a little deeper here, we are starting to see tightly defined automation projects stall, when the true chaos of what lies beneath the surface is exposed. Processes and business activities that management thought were solved problems in need of a bit of tidying up can turn out to be convoluted human work-arounds with a labyrinthian number of alternative options to reach business closure. Similarly, using task and process mining tools to contrast ‘best practices’ is sometimes misused by enterprises under the impression that the fast route is the best route. For example, Worker A takes 20 steps and 4 hours to complete a task, whereas worker B takes 12 steps and ends the same task in 30 minutes. Often, worker B is designated the one to model and for everyone (bot or human) in the future to take this route. But common sense, which is not, in reality, all that common, tells us that worker B may well be cutting corners and potentially doing a poor and incomplete job.

KMWorld Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues