When automation is not the right move — and how to spot it early
Automation conversations tend to focus on what to automate. Less often do they address what not to automate — which is usually the more important question.
Automation applied at the wrong time or to the wrong process does not save time. It shifts problems onto a machine, adds maintenance work, and makes it harder to improve the process later.
Sign 1: The process changes frequently
If the process changes month to month — steps shift, exceptions accumulate, logic evolves — automation creates more friction than it removes. Every change to the process requires a corresponding change to the automation. At best, that is a few hours of work; at worst, a full rebuild.
Automation's value grows when the process is stable. Before automating, let the process mature: run it manually, sharpen the steps until they are clear, and surface the edge cases. Automating at that point produces real returns.
A practical test: has the same process stayed unchanged for the past three months? If not, it is not ready to automate.
Sign 2: The process depends heavily on judgement
Automation is good at executing rules — not at exercising judgement at their edges. If the critical parts of a process require situational assessment — "in this case we make an exception" or "this needs a human review before we proceed" — automation either blindly skips those moments or stops and waits for a person.
Both are costly. The first produces errors; the second removes the benefit of automating at all.
The answer is not always to add more logic. Sometimes the better question is: can the process be simplified so that judgement calls become rare? That makes the process better even without automation.
Sign 3: Volume is too low for the automation to pay for itself
Automation requires time to build, time to test, and ongoing maintenance. If a process runs once a week or less and takes only a few minutes, the payback threshold is high.
A simple calculation: how many hours per year does this process consume? Multiply by the hourly cost of the person doing it. Is that figure larger than the investment needed to build the automation?
At low volumes, the answer is often no. In those cases, a better solution may be to simplify the process or create a clear template that is followed manually. Both are faster to implement and lighter to maintain.
Sign 4: The process is not documented
If nobody can describe the process step by step before automation begins, automation will not help. Automation executes the process as it is described — not as you wish it were.
In this situation, documentation comes before automation. When the process is written out, the unclear or redundant steps usually become visible immediately. The process improves before a single line of code is written.
Sign 5: The problem is actually organisational
Sometimes what looks like a process problem is really a communication problem, an unclear ownership problem, or a missing piece of information. Automation does not fix these — it automates the symptom, not the cause.
An example: an ordering process is slow because approval always waits for one person who is frequently unreachable. Automation can send reminders, but the real solution is a substitute arrangement or expanded authorisation.
What to do instead
When automation is not the right solution right now, that does not mean it never will be. In many cases, the best path forward is:
- Document the process as it currently runs
- Simplify or stabilise the steps
- Identify which parts repeat most often and change least
- Automate those parts first
The smallest viable automation is better than a broad one that tries to cover everything. And a process that runs well manually is always a better starting point than a chaotic one that is merely fast.
An honest assessment before starting
An automation consultant's job is not to sell automation. It is to help assess when automation makes sense — and when it does not. If an assessment concludes that automation is not the right step now, that is a valuable result, not a failure.
Get in touch — I can review your process and help you decide whether automation is the right next step or whether something else would serve you better.