Problem
Automation has become the default answer to any annoyance in a business.
Something feels fiddly? Automate it. Someone makes mistakes? Automate it. You’re busy? Automate it.
The result: a pile of half-working zaps, brittle spreadsheets, and “systems” that create just as much maintenance as the problem they were meant to solve.
Automation isn’t the prize. It’s a tool. And like any tool, it should earn its place.
Principle
Automate something when it clearly does at least one of these jobs:
- Saves meaningful time or money
- Protects or improves quality
- Creates a better customer experience
If it doesn’t do one of those, you’re probably automating for the feeling of progress - not the outcome.
A useful check
Before you automate anything, be able to finish this sentence:
“If we automate this, it will _____.”
If the answer is vague (“make things smoother”, “help us scale”) you’re not ready. You need a more specific outcome.
Examples (what this looks like in real life)
1) Saves time or money
Good automation:
- Automatically create an invoice draft when a project hits a milestone.
- Turn a form submission into a task with due dates and a checklist.
Not worth it (yet):
- Spending 6 hours automating a task you do once a month for 3 minutes.
Rule of thumb: if the time saved is tiny, your automation won’t pay rent.
2) Protects or improves quality
Good automation:
- Prevent missed steps (e.g., onboarding checklist auto-created every time someone buys).
- Standardise handoffs (e.g., every new lead gets tagged, logged, and routed consistently).
Not worth it (yet):
- Automating something that’s messy because the process is unclear. You’ll just automate the mess.
If people are making mistakes, ask: is the process unclear, or are they just human? Automation is great at consistency - terrible at decision-making in a fog.
3) Creates a better customer experience
Good automation:
- Send a clear “here’s what happens next” email immediately after purchase.
- Confirm bookings instantly with the right links and expectations.
- Route enquiries so they don’t sit in a void for three days.
Not worth it (yet):
- Automated sequences that add noise, not clarity.
The bar here is simple: does the automation reduce customer uncertainty or effort?
Practical advice (how to decide fast)
Use this quick decision filter on any automation idea:
Step 1 - What’s the job?
Pick one primary goal:
- Save time/money
- Improve quality
- Improve customer experience
If it needs “all three” to justify it, it’s usually not clear enough.
Step 2 - What breaks if it fails?
If the automation fails, will it:
- cost you money?
- create customer confusion?
- cause data loss?
- silently drop something important?
If failure would be expensive, start simple and add monitoring/alerts before you get fancy.
Step 3 - Is the process stable?
If the steps change weekly, don’t automate yet. Document the steps once, run them a few times, then automate the stable bits.
Step 4 - Automate the boring middle, not the thinking
Automate:
- capture
- routing
- reminders
- formatting
- repeatable actions
Don’t automate:
- judgement calls
- nuanced communication
- “it depends” decisions
Your Action
Pick one repetitive process you do in the business (or one that regularly gets dropped).
Run it through these three criteria:
- Will it save meaningful time or money?
- Will it protect or improve quality?
- Will it create a better customer experience?
If the answer is “yes” to at least one, write down the specific outcome you want (one sentence), then automate only the smallest part that achieves that outcome.
If the answer is “no”, you’ve just saved yourself a bunch of busywork disguised as productivity.