Everyone Wants Automated Processes. Almost Nobody Has Anything to Automate Yet.
Do your automation projects stall out in chaos or never quite finish? After running operations and technology projects for service businesses across more than 70 client engagements, I’ve seen the same pattern often enough to name it — and once you see it, you can’t unsee it.
A client comes to us and says: we want to be digital, we want to use AI, we want automated processes. The energy is real and the intent is right. So we dive into the engagement, go looking for the processes to automate — and find there’s nothing there to work with.
Not because the business is broken or the team is incapable, but because automation is the last stage of a maturity curve, and they’re trying to skip the ones that come before it. Skip those steps and one of two things happens:
There’s nothing concrete to automate at all, or you end up automating something so loosely defined that the results are unreliable, brittle, and quietly worse than the manual version they replaced.
That realization — that process has a maturity to it, and automation sits at the very top — came out of doing this work across dozens of businesses, not from a framework I read somewhere. Here’s the scale as we’ve come to see it, and why the real work almost never happens where people expect.
The five stages of process maturity
Stage one: no processes at all.
The typical early-stage startup. Things get done, and often done well, but there’s no repeatable way they get done. Every project is improvised. This is fine at the very beginning — it’s actually a reasonable way to move fast when you’re tiny — but it doesn’t survive growth.
Stage two: processes in people’s heads.
There genuinely is a way things run — it just lives in the founder’s head, or is scattered across the team, with each person holding their own slightly different version. Nothing is written down. This is where a huge share of service businesses actually sit, including ones doing serious revenue. It works right up until a key person is on holiday, leaves, or the team grows past the point where tribal knowledge can hold it together.
Stage three: documented processes.
For the first time, the process exists outside anyone’s head — SOPs, flowcharts, a standardized, agreed way of doing the work that anyone can point to and follow. This is the first stage where the process belongs to the business rather than to individuals.
Stage four: digitally supported processes.
Now you take that defined, standardized process and start augmenting it with tools. Think Monday.com boards, templates, structured workflows in whatever software you already run. The process is still executed by people, but software now scaffolds it — enforcing steps, catching handoffs, reducing the load on memory.
Stage five: automated processes.
Only once you’re solidly at stage four can you meaningfully move here. This is the stage everyone names when they first call us. It’s the destination — but it’s a destination you reach by climbing, not by teleporting.
The gap that actually matters is between two and three
Everyone wants to be at stage five. But the reality — especially in service businesses — is that most companies sit at one or two. And here’s the part worth sitting with: so do most of your competitors.
That completely changes where the leverage is. If your rivals are also stuck at “it’s all in our heads,” you don’t need automation to pull ahead of them. You need to get to documented and aligned faster than they do.
Across the 70-plus engagements I’ve run, the advantage has consistently come not from the flashy stage-five work, but from the disciplined climb out of stage two. The real work — the work that creates a durable operational edge — happens in the move from stage two to stage three.
That’s the unglamorous part nobody’s excited about. It isn’t AI and it doesn’t demo well. It’s sitting down — internally with your team, or with someone brought in to run the process — and doing two specific things: getting genuine alignment on how your processes actually run today, and documenting them so they’re captured for good. Both matter, and they’re not the same thing.
Alignment without documentation evaporates the moment the meeting ends. Documentation without alignment is fiction — a tidy diagram of a process nobody actually follows. You need both, together, before anything you build on top of them will hold.
In practice this looks like mapping a process with the people who really do the work, surfacing the places where three team members each thought a different person owned a step, writing the SOP that resolves it, and standardizing on one agreed version. Undramatic. Also the single highest-return thing most service businesses can do.
Why you can’t skip ahead
The reason this sequence isn’t optional comes down to a simple mechanical fact: as long as a process is undefined, it varies with every execution. And you cannot meaningfully automate something that changes every time you run it.
Automation needs a stable, repeatable input — the same steps, the same triggers, the same handoffs, every time. An undefined process doesn’t give it one, so you either can’t automate it, or you automate a snapshot that’s wrong half the time it runs.
So when a client wants to jump straight to stage five, we’re not slowing them down out of caution or process for its own sake. We’re building the thing that makes automation actually work later.
Get the process defined and standardized at stage three, augment it with the right tools at stage four, and then automation has something solid to run on — and it produces results worth having, rather than a fragile system that breaks the first time reality doesn’t match the diagram.
If you’re eyeing AI and automation for your operations, the honest first question isn’t “which tool should we use?” It’s “where are we on this scale, really?”
For most service businesses, the answer is stage one or two — and the highest-return move on the board isn’t the automation everyone’s excited about. It’s getting to stage three.
