I'm seeing a significant amount of conflation with these two things, and as someone who has made a career out of business process improvement through the implementation of software and automation, I can tell you that we've been doing automation (successfully) without AI for a very long time.

Before writing a line of code, choosing an automation tool or (PLEASE DEAR GOD) asking an LLM to do your thinking for you, there is a way to look at business processes to identify what, if anything, needs intervention.

When it comes to your business, there are three dials that you can turn to see improvement, and they are labeled:

"People"
"Processes"
"Products"

These three dials, if adjusted or changed without consideration, can turn simple business change into a three body problem faster than you can say "Henri Poincaré."

Here’s the simplest way I know to explain it: every business is a system, and every system has a limiting factor.

That idea sits at the heart of something called the "Theory of Constraints", but you don’t need to know the theory to recognize it in practice. You’ve felt it any time you worked harder, moved faster, or bought better tools—and nothing really improved. Work just piled up somewhere else.

That’s because systems don’t get better when everything gets faster. They get better when the slowest, most fragile, or most confusing part gets attention. Speeding up everything except the bottleneck doesn’t fix the system. It just creates more pressure where things already struggle.

This is where automation and AI so often go wrong. They’re applied broadly, enthusiastically, and early—before anyone has paused to ask where work actually slows down, where decisions stall, or where humans are forced to compensate for unclear processes.

You can see this clearly in real estate.

From the outside, it’s tempting to think the problem is follow-up. So people reach for tools that automate emails, texts, and reminders. But if you zoom out and trace the actual flow of work—lead comes in, conversation happens, paperwork starts, timelines emerge, inspections happen, documents move, decisions get made—the friction usually isn’t writing the next message. It’s visibility. It’s not knowing, at a glance, where a deal actually stands or what the next real constraint is.

Adding smarter messaging doesn’t fix a broken hand-off between stages. Automating lead capture doesn’t help if nothing connects cleanly to transaction management. The system doesn’t need to be louder or faster—it needs to be clearer.

The same pattern shows up in small consulting businesses.

These are often smart, capable people who are constantly busy. Work arrives through email, LinkedIn, referrals, DMs. Project details live in proposals, shared docs, Slack threads, and someone’s head. When a client asks for an update, the consultant has to stop, search, reconstruct context, and reorient themselves.

The issue isn’t effort or expertise. It’s not even scale. It’s that the system requires constant thinking just to know what’s going on. That’s the constraint. Until that’s addressed, adding automation—or AI—just accelerates confusion.

This is why I think about business improvement as three dials: people, processes, and products.

You can turn any one of them. Hire more people. Change how work flows. Buy new tools. But if you turn the wrong dial first, or turn one without considering the others, you can make a simple change feel like chaos almost instantly.

Well-designed automation tends to feel boring. Calming. Quietly helpful.

Poorly designed automation feels like speed without control.

AI is powerful. Automation is powerful. But neither replaces the human effort of understanding how work actually moves through a system, where it slows down, and why.

That part still requires thinking.