In the early 1990s, executives and managers welcomed information technology — databases, PC workstations, and automated systems — into their offices. They saw the potential for significant business gains. Computers wouldn’t just speed up processes or automate certain tasks — they could upset nearly all business processes and allow executives to rethink operations from the ground up. And so the reengineering movement was born.
Now it’s happening again. Powerful machine-learning algorithms that adapt through experience and evolve in intelligence with exposure to data are driving changes in businesses that would have been impossible to imagine just five years ago. The PCs and databases introduced during the reengineering of the 90s have grown up: the rules-based codes written by engineers are giving way to learning algorithms driven by the machines themselves. As a result, business processes are being machine-reengineered.
Algorithms aim to redesign business processes just like humans did during the original reengineering movement. Then, reengineering was limited by the speed of humans. Managers noted historical trends and revised processes, and engineers developed code that was then baked into computing systems. Every update or response to the market required multiple steps; it cost time and performance. Sometimes, by the time changes were in place, the market had already moved. With machine-reengineering, process changes are constant and driven not just by history but also by the predictive capabilities of machine-learning algorithms. Machine-reengineering asks that people train and actively manage the performance of the algorithms and data models that drive process change, rather than drive process change themselves.
Reengineering got off track by encouraging businesses to overhaul too many processes too quickly. Moreover, the reengineering rhetoric of “obliterate” was extreme and ultimately destructive not only to processes, but to businesses as well. Machine-reengineering seems to have so far avoided these mistakes. Businesses that machine-reengineer their processes focus on one core process at a time, and thus they can quantify positive outcomes.
In our study of more than 30 pilots in early-adopter companies, we found five common business processes improved by machine-reengineering. The chart shows the proportion of companies targeting a business process improvement category, and then the proportion of those processes that used various machine-learning techniques to accomplish the goal. For example, nearly half of the companies are using machine reengineering for marketing products and services. If you hover over that purple bar, you see they’re doing this predominantly through predictive analytics, but also mixing in some visual sensing and natural language detection.
Once we mapped the processes targeted to the machine-learning techniques used, we wanted to understand how those techniques connected to three desired outcomes for the business: improving cost performance, customer performance, and revenue performance. For example, hovering over natural language processing’s orange bar shows that it’s creating improvements in cost and customer performance in near equal measure, but it’s not really applied to revenue performance.
That’s the proportional view of activities. How is all this machine reengineering actually paying off? Though this is just the beginning (we suspect many more processes will follow) we already see evidence of significant, even exponential, business gains in these three areas.
Nearly half of early movers reported improvements to top-line performance. Most often, improvements came through automatically providing more timely predictive data to employees who interact with customers or sales prospects.
