The Great Disconnect
Automation is sold as a "magic-button" to the Executive class. So why do so many implementation projects fail or stall?
Never trust a salesman, always ask the technical person! (same holds true for buying a car, talk to the mechanic).
Executives overestimate automation because they don’t truly understand the manual nuances of the processes they are trying to replace. A fundamental knowledge gap exists between the C-suite, middle management, and the front-line workers.
The "Autopilot" Problem: What Employees Don’t Tell You
Employees perform dozens of "unwritten" micro-tasks every day that never make it into a job description. Many tasks are done on autopilot (e.g., a quick keyboard shortcut to fix formatting or a mental check of a secondary software).
Example: Creating a spreadsheet isn't just "moving data"; it involves verifying sources, cleaning messy inputs, and cross-referencing emails. (steps the employee might not even realize they are doing.)
Not to mention all of the things we do throughout the day that we only register subconsciously:
- You pass by Jan in the hallway and recognize she has a lackadaisical demeanor.
- You pass by John and see that he has a scowl on his face.
- You sit down at your desk and decide to move the 1:30 pm meeting to next week. You just think today isn't a good day for it. But you didn't actively consider why that is.
- Hint, it has something to do with Jan and John.
The Layers of Ignorance (The Hierarchy Gap)
Managers have a "vague idea" of the workflow but lack the granular detail of the day-to-day struggle. Their job is to manage the big picture and the employees are the ones who actually know how to get their specific part of the process done.
Executives are even further removed. They see the output (the finished product) but are almost entirely disconnected from the "nuances and intricacies" of the production line.
The employees in turn, have no idea what it is like to see the workflow of the business from the top. The ppl on the factory floor don't know what it's like to deal with the Global Supply Chain responsible for shipping the product they've made.
The Results of not doing proper "Discovery":
- Leadership buys software for a process that exists in their head, not the one that exists in reality.
- Employees can't use the new the automated business process, because it's missing dozens of edge case items the execs never knew about.
- Half your customers become irate because Susan knew exactly what they were trying to say, whereas this new "Robot" can't make sense of anything.
Not doing proper discovery kills Automation ROI
If you automate a process without accounting for the "hidden steps," the automation will break the first time it hits an edge case.
The Human Element:
Since LLMs hit the market, there's been doomers talking about eliminating whole workforce(s).
Mass replacement won’t work. This isn't a moral or ethical argument. It’s a practical one. Total automation fails because it ignores the "Human Layer" that keeps a business from falling apart.
- Un-Quantifiable Data: Human communication and action happen on a level a machine simply cannot comprehend. We process thousands of tiny details every second:
- the tone of a voice
- the subtext of an email
- the "vibe" of a room
- The "Susan" Factor: Remember Susan? She’s the one who knows that when a client says "it’s fine", it actually means "I’m about to cancel my contract." A machine hears "fine" and moves to the next task. Susan hears the subtext and saves the account.
- Contextual Judgment: You can't program a robot to know that today is a bad day for a 1:30 PM meeting because Jan is stressed and John is frustrated. Humans act on "subconscious data" that hasn't been (and often can't be) digitized.
Discovery is the Most Difficult Part of Automation
In order to automate a process, I first have to understand the process. In order to do that, your workforce has to understand their process.
Essentially people pay me to document their workflow first, and automate it second.
Automation is easy, documenting a workflow is hard.
I've turned processes that took a week into a process that takes a day. But it took month(s) of research before I even started coding.
Before buying the tech, do a "deep dive" or process mapping session with the people actually doing the work.
Final Thought: Respect the complexity of "simple" tasks.
We need to stop calling manual tasks "simple." If a task requires human judgment, it is inherently complex. Respect the intricacies of the work your people do. Because you can’t automate what you don’t understand.