In 1896, the American architect Louis Sullivan wrote that "form ever follows function." He was talking about buildings, meaning that shape should be a consequence of purpose.
What’s Behind the Curtain
Local BPO legend has it that a handful of VC-backed "business automation" tools, machine-forward, just add data from a convenient app, were, in fact, teams of data encoders working in converted warehouses on the former Clark Air Force Base.
These were pre-AI and branded as automation.
Users got their outputs. Rigorous processes and a high body count ensured quality data: calendars were booked, reports were timely... Most customers, confronted with evidence of the humans behind the curtain, didn’t really care.
Lesson learned: people love an automation story.
Lesson lost: if it works, it works.
Two decades later, and we're still in the throes of tool infatuation. If you've ever tried to automate a report with internal calculations using Claude or ChatGPT, you know exactly what I'm talking about.
The Tool Is Not the (Whole) Answer
Sullivan's principle, updated. Operators who open an execution problem constrained by their tools limit their options.
Sometimes you want a hollow chisel mortiser for fast, repeatable production work. Sometimes you want a Japanese bench chisel and mallet when you're working with dense or unpredictable hardwood, where a mortiser would blow out the grain. Sometimes you want a router with a spiral bit to rough out the waste, then pare the walls clean by hand.
You have to know what you’re doing and work backward.
The Wizard of Oz Was Right
Pay no attention to the man behind the curtain.
Those automated tools ran on hidden human labor — and they worked. When they scaled, the data returned clean, structured, accountable.
When those companies failed, it wasn't because the humans were discovered. It was because the output broke: latency, buggy upgrades, lousy support.
Accountability to outcome is the point.
Scope (Function) = Execution Variable
Most execution failures — missed deadlines, quality problems, rework cycles — trace back to the same root: undefined define work, undefined function.
Vague briefs produce vague outputs. "Social posting" produces content that may or may not match brand voice, may or may not go through approval, may or may not meet deadlines.
Form follows function. Task type. Volume. Format. Turnaround time. First-pass acceptance rate. Escalation path when something is wrong.
Brief → Assign → Deliver
Before execution, define outcome. The first step is scope confirmation: service area, task volume, acceptance criteria, what good looks like.
The operators who got the most from those Wizard-of-Oz platforms weren't the ones who understood the back office. They were the ones who knew exactly what output their customers needed. That clarity is what made the hidden human labor legible, scalable, and eventually — in the better companies — replaceable with actual automation as the technology evolved.
Define the outcome well enough, and the method can change seamlessly
