Field notes / Automation
Automation can make a good process dramatically better. It can also make a bad process fail faster.
Igniter Studio · 2 Oct 2026 · 4 min read
Automation is attractive because the promise is easy to understand:
Do less repetitive work.
Move faster.
Make fewer mistakes.
Keep systems in sync.
All good things.
But automation has one uncomfortable characteristic.
It is very good at repeating exactly what you tell it to do.
Including bad processes.
If a workflow is confusing, unnecessary or badly designed, automation does not necessarily fix it.
Sometimes it simply performs the confusion faster.
Before choosing an automation platform, map what actually happens.
Not what is supposed to happen.
What actually happens.
For example:
Now ask why each step exists.
Maybe the CRM can receive the form directly.
Maybe assignment can be based on location.
Maybe the acknowledgement email does not require a person.
Maybe reminders can be generated automatically.
Maybe one of the systems is completely unnecessary.
Understanding the workflow often produces improvements before a single automation is built.
Good candidates usually share a few properties.
They happen frequently.
The steps are predictable.
The inputs are reasonably structured.
Mistakes are annoying or expensive.
And a human is currently doing work that does not require much judgement.
Examples include:
None of this is glamorous.
That is precisely why it is worth looking at.
A process can contain both mechanical and judgement-heavy steps.
You do not have to automate the entire thing.
Consider recruitment.
Software might:
That does not mean software should make the final hiring decision.
The same applies elsewhere.
Automation is strongest when it handles predictable transitions and gives people better information for the decisions that remain.
The goal is not to remove humans from the workflow.
It is to remove work humans should not have been doing manually in the first place.
An automation that works 90% of the time can be worse than no automation.
Because now people need to trust it while also checking whether it failed.
Good automation should consider failure explicitly.
What happens if an API is unavailable?
What happens if required data is missing?
What happens if the same event arrives twice?
What happens if an email fails?
Can somebody see what happened?
Can the action be retried safely?
These details are not exciting, but they separate useful automation from brittle demos.
One danger with highly automated systems is that nobody knows why something happened.
A customer receives an email.
A record changes status.
A task disappears.
A report arrives.
But the team cannot easily explain what triggered it.
That becomes painful when something goes wrong.
Important automations should leave useful traces.
That might mean:
People should not need to call the developer to understand whether the system worked.
Businesses often imagine automation as one enormous transformation project.
It does not need to be.
Fixing five small recurring processes can produce more value than trying to automate an entire department at once.
Consider:
Monday: automatically prepare the weekly pipeline summary.
Tuesday: sync new enquiries into the CRM.
Wednesday: remind customers with incomplete onboarding.
Thursday: organise uploaded documents correctly.
Friday: prepare management reporting.
Each improvement may save only a little time.
Together they change how the week feels.
An automation should reduce operational burden.
If it needs frequent manual repair, it may simply have moved the work somewhere less visible.
This happens particularly when workflows depend on fragile browser interactions, unofficial integrations or assumptions about external systems.
Whenever possible, use stable APIs, webhooks and documented interfaces.
The boring implementation is often the reliable implementation.
That is a good trade.
You do not need an “automation strategy” to begin.
Find something people complain about repeatedly.
Something like:
Every Friday I have to download these files, combine them and email the result.
Or:
Whenever we get an enquiry, I copy the details into three places.
Or:
We keep forgetting to follow up after seven days.
Map it.
Simplify it.
Then automate the predictable parts.
Once that works reliably, find the next one.
The best automation programmes are often just a collection of small sensible decisions.
No robot revolution required.
Your next chapter starts here
Tell me what you’re working on. I’ll help you find a practical way forward.
Let’s talk about it