When a business gets busy, the natural response is to add something. Another person. Another tool. Another meeting. Another spreadsheet. Sometimes that is exactly what the company needs. But I have also seen the opposite: every addition creates another hand-off, another place to look and another thing that somebody has to remember.
I prefer to start one level lower. What is the work that actually needs to happen? What enters the process? What decision has to be made? What information is required? What should happen next? Who owns the result? Those questions are not exciting, but they expose a surprising amount of waste.
A lot of complexity is really uncertainty wearing a suit. If nobody knows which leads are worth pursuing, the company creates more reports. If nobody knows who owns a customer problem, more people get copied into messages. If a task is repeatedly forgotten, another meeting appears. The organization becomes heavier without becoming more capable.
The first useful system is often extremely simple. One clear intake. One definition of status. One owner. One next action. One place where the result is recorded. That is enough to turn a recurring problem into something that can be improved.
Only after the process is visible does automation make sense. Then software can remove repetitive steps without hiding the logic. A form can structure the input. A rule can route the work. An AI model can classify or summarize. A notification can arrive when a threshold is crossed. The technology becomes leverage because the business already knows what it wants the system to do.
This matters even more when hiring. People are expensive, but the bigger cost is often putting good people into bad processes. If somebody spends half the day searching for information, chasing approvals or correcting avoidable mistakes, hiring another person does not solve the operating problem. It just increases the number of people living inside it.
The same is true for software. A company can subscribe to excellent products and still have a poor system. Tools are components. The operating model is the way those components, people and decisions work together. Buying more components does not automatically produce a better machine.
I am not arguing for endless process design before doing anything. The opposite, actually. The fastest way to build a useful system is to work close to reality. Run the process. See where it breaks. Fix the recurring failure. Write down the rule. Measure the outcome. Repeat. A system should grow out of real work, not a diagram made in isolation.
There is a point where a process becomes stable enough that people can trust it. That is an important moment. The company stops depending on heroic memory. New people can understand how work moves. Management can see where something is stuck. Automation can be added without guessing. Improvement becomes cumulative instead of starting from zero every week.
That is what I mean when I say build systems before adding complexity. Complexity is not always bad. A serious business will eventually have plenty of it. The goal is to earn that complexity by making each layer useful, understandable and accountable.