Where to start with automation
The hardest part of automation is not building it, it is choosing what to automate first. Pick the wrong task and you burn effort for little return. Here is how to find the opportunities that pay back quickly and build momentum.
How to spot a good automation candidate
The best first automations share a shape. They are repetitive, so the saving happens again and again. They follow rules or patterns, so a machine can learn them. They happen often or at awkward hours, so a person doing them is either expensive or absent. And they are error prone when done by hand, so automating them improves quality as well as speed. When a task ticks several of those boxes, it is worth a serious look.
The tasks to avoid at the start are the rare, judgement heavy ones where every case is different. They are the hardest to automate and the slowest to pay back. Start where the work is dull, frequent and predictable, and save the clever edge cases for later.
A simple way to prioritise
Repetitive
The more often a task runs, the more a small saving per run adds up across a month and a year.
Rules based
If you can describe the task as a set of steps or rules, it is a strong candidate for automation.
High volume or after hours
Work that floods in, or lands when nobody is there, is where automation earns the most.
Error prone by hand
If manual work causes mistakes, automating it improves accuracy as well as saving time.
Clear owner
Someone who feels the pain and will help embed the change makes the difference between success and shelfware.
Value over effort
Rank candidates by the payback against the effort to build. Start where that ratio is strongest.
Common questions
Should we automate the biggest problem first?
Not always. A slightly smaller problem that is quick to automate and pays back fast often beats a huge one that takes months and carries more risk.
Is this automation or AI?
Both. Simple rules based work suits classic automation, while messy language or decisions suit AI agents. The right tool depends on the task.
What about the exceptions?
Exceptions are where automation usually breaks. A good design handles the routine automatically and routes the odd case to a person cleanly.
How do we avoid automating a bad process?
Look at the process first. Sometimes the win is fixing or removing a step, not speeding up a task that should not exist.
More from our consultancy
Not sure what to automate first?
Book a short session and we will help you rank your opportunities and pick a first project that pays back fast.