In Part One of our AI-Readiness series, we explored Information as the foundational element of automation and Artificial Intelligence (AI). Part Two of this series focuses on answering the burning question on the minds of business leaders taking the leap into the world of AI: "What should I know before I get started?"
Process Automation, which includes Robotic Process Automation (RPA), is a crucial area that tricks many well-meaning people into the fallacy of an AI promised land. Unfortunately, process automation is touted as a fool-proof, flip-the-switch solution that magically produces cost takeout, efficiency, and reduces or replaces human efforts. While those benefits are attainable through automation, and eventually AI solutions, the work to be done to gain value and optimize quickly is more than currently discussed. However, don't worry! Our promise to you as part of this Readiness series is to help you navigate these waters with eyes wide open. You will benefit from our time with other clients who have ventured in and lived to tell the tale--even if it was after a few scraped knees from lessons learned the hard way.
Teaching Machines What You Know
If you want to use AI to improve processes, the most critical step is capturing tacit knowledge. There are things that employees know from their day-to-day work that need to be extracted and documented to enable the machines to work correctly. Even if you have well-documented procedures, there is undocumented information about how the work gets done and how people make decisions. Human experience and context are often non-documented drivers of efficient and effective processes. Machines and bots live by rules and only do what you tell them; this information must be captured and added to current process flows for the technology to work "like a human."
This collection process is likely either annoying or frightening news to you. The day-to- day work needs to be translated into a set of rules and steps for machines to do the work that comes naturally to humans. Collecting this information is work that takes time. It may feel like it contradicts the goal of automation speed and efficiency. Yet, it’s the only way you can build solutions that work. Otherwise, you end up with solutions that can't quite get the job done.
In context, you should be aware that not knowing exactly how your team gets their work done is not a leadership gap. It is not efficient for you to know everything. The beauty of human beings is their ability to adapt without having to be told to--unlike RPA. Your responsibility here is to uncover the ways humans have naturally adapted to their work so you can tell the machines to follow the same process.
Three Key Steps to Capturing Tacit Knowledge
Invest time in capturing and documenting processes that will help you create a comprehensive, intelligent solution
- Schedule one-to-one interviews or small focus group sessions to review processes that are being automated. During the session, capture information gaps, processes, and decision steps.
- Conduct ethnographic research to observe employees performing their job duties for a half-day. This will help capture any missing steps in the process that weren’t articulated during the one-to-one interviews. Task those performing the observation to weigh in on gaps, what's not written down, and how they make decisions at key process points. We do things without realizing it, so we can't always articulate everything when interviewed.
- Finally, write down what you observed-in detail. Tacit knowledge remains elusive if it is not explicitly recorded, reviewed, and revised so that it’s understandable by a machine.
If you invest this human investigation time up front, you will be sure to create a comprehensive, intelligent solution. There will be less time spent "fixing" if you take the time to get it right proactively.
As usual, if you genuinely want to maximize all that these solutions have to offer, there is a human component to be addressed and, in this case, captured. An AI-Ready Process helps you rethink how you improve process effectiveness and efficiency. While you can enhance efficiency, you can't impact effectiveness without uncovering the un-documented details on how the work gets done.
Up next, in Part Three of this series, we will explore AI-Ready Culture.