"Automation" sounds like robots and artificial intelligence. In most businesses it is far more ordinary: removing the manual steps that eat hours every week. Done well, it is one of the most reliable ways to give a small team its time back.
What automation actually means
At its simplest, automation means a computer does a step that a person used to do: sending a reminder, moving information from one system to another, generating a report, routing a request to the right person. No magic, just fewer things falling through the cracks.
What makes a good candidate
Look for work that is:
- Repetitive. The same steps, again and again.
- Rules-based. Decisions that follow clear criteria.
- High-volume. Done many times a week.
- Error-prone. Mistakes creep in when people copy and paste.
- Waiting-heavy. Things sit in someone's inbox until they remember.
Good first automations
- Approvals. Requests routed automatically, with reminders when they stall.
- Notifications. Customers or staff told when something changes, without anyone remembering to tell them.
- Data entry between systems. Information entered once and passed along.
- Reports. Weekly numbers assembled automatically instead of by hand.
- Onboarding. Checklists that create tasks for each person involved.
- Chasing. Gentle automatic reminders for overdue invoices or missing documents.
How to choose
List candidates, then score each on time saved per week, error cost, and how clearly the rules can be written down. Start with something that scores well on all three and is not mission-critical, so you can learn safely.
Simplify before you automate
Automating a messy process produces a faster mess. Walk through the process first. Which steps exist only because "we have always done it"? Remove those, then automate what remains.
Start small and measure
Pick one process. Automate it. Measure the time saved and the errors avoided. Use that to decide what to do next. Small, visible wins build trust in the approach.
Keep a human in the loop for exceptions
The best automation handles the common case and flags the unusual one for a person. Aim for a process that runs itself 90% of the time and asks for help the rest, rather than one that is rigid and breaks.
Common mistakes
- Automating before understanding the process.
- Building something so complex nobody can maintain it.
- Forgetting to tell the team what changed.
- Ignoring failures. Automations need monitoring: what happens when a step fails?
Where software comes in
Some automation can be done with existing tools. When it spans several systems or follows rules specific to your business, a small custom workflow is often cleaner. Related reading: connecting your business systems.
Not sure where to start? Describe the process that frustrates you most. It is usually a good place to begin. You can also see what we build.