Inefficient at the speed of light

Moore’s law is the gift that keeps on giving. It has enabled countless breakthroughs as we’ve passed from the Industrial Age to the Information Age to the Knowledge Age. More recently, its impact has been felt in areas such as text and data analytics, machine learning, and, more recently, generative AI. But it can also be a curse. As CPU speeds, memory capacity, and bandwidth keep ratcheting upward, it’s easy to get complacent—and sloppy. No place is this more evident than in the countless business processes that are deeply entrenched in nearly, if not every, function within an enterprise. But massive amounts of “make work”—paper shuffling, approval cycles, check boxes, and the like—have no place in the enterprise of the future. The question is, what can we do about it?

We keep hearing about the breathtaking gains in speed and efficiency brought about by AI-enabled hyperautomation. But how much of that automation is actually intelligent? Unfortunately, much of today’s “digital transformation” isn’t really transformation at all—it’s just putting old wine in sleek, new, synthetic wineskins. And much of that old wine consists of shuffling mountains of digital paper generated by long-standing legacy policies and procedures. Now would be a good time to step back, peer under the hood, and do some cleaning and trash collecting. Fortunately, we have just the right technology to accomplish that. Welcome to the dull, boring, watching-paint-dry world of … process mining.

Process mining has been around since 1998, when researchers began analyzing workflow logs and event-based data to discover implicit process models. This work ultimately resulted in the IEEE 1849-2023 XES (eXtensible Event Stream) XML standard for capturing and analyzing event data. Building upon this standard, process mining combines the interchange of event data among information systems in multiple application domains with state-of-the-art data analytics. With an annual growth rate of around 40%, the size of the process mining market is expected to reach $10 billion by 2030. And yes, Gartner has a Magic Quadrant for process mining platforms that you can check out. The potential performance gains are significant. According to Deloitte, in a 2019 Harvard Business Review article (hbr.org/sponsored/2019/12/ how-companies-are-using-intelligent-automation- to-be-more-innovative), when properly designed and implemented, intelligent automation can result in a threefold boost in annual revenue growth. When combined with an average cost reduction of 21%, greater overall performance and profitability result.

Have you ever wandered through one of those corn mazes that pop up during autumn, especially around Halloween? They can be fun for children and adults of all ages. But hard as it is to imagine, you are still likely to encounter the not-so-fun experience of wandering through a similarly endless click maze when you find yourself trapped in a phone- or web-based help desk. It may be a logical, well-structured process, but with little or no thought given to the end-user experience. How many times have you made the effort to carefully scroll through an FAQ, and, not having found anything close to what you were looking for, entered your question in the search box, only to be told by a chatbot to go check the FAQ? Clearly, there’s a disconnect. Process-driven thinking has overtaken customer-centered design. Let’s take a look at all of this through a KM lens. But, before we do, we need to ensure that our own KM process house is in order.