The easiest way to waste time with automation is to start with the technology.
A new AI tool appears. A platform launches a new feature. Somebody says every business should automate everything.
The more useful starting point is much less exciting:
What do we repeatedly do that follows a predictable process?
Find repetition first
Look at a normal week.
Which tasks happen again and again?
Examples might include:
- sending confirmations
- creating customer records
- chasing missing information
- updating statuses
- producing invoices
- reminding customers
- moving information between systems
- notifying staff
- preparing standard reports
- following up enquiries
Write them down before choosing any software.
Prioritise by frequency and friction
A task that takes two minutes but happens 200 times a month may be more valuable to automate than a two-hour task that happens twice a year.
I normally think about four factors:
- Frequency — how often does it happen?
- Time — how much human time does it consume?
- Consistency — does it follow clear rules?
- Risk — what happens if the automation gets it wrong?
High-frequency, rules-based, low-risk tasks are usually strong early candidates.
Connect the workflow, not just one action
Sending an automatic email is useful.
But the bigger opportunity is often connecting the actions around it.
For example, a new booking might create a customer record, assign the booking, collect payment, update capacity, send the correct information and notify a staff member.
That removes several manual steps instead of one.
Do not automate broken processes
If a workflow is confusing when humans run it, automating it can make the confusion happen faster.
Simplify first.
Decide what the correct process should be, where exceptions occur and who owns the outcome.
Then automate the stable parts.
Use AI where judgement is helpful, not where certainty is required
AI can be extremely useful for drafting, summarising, classifying and assisting with decisions.
But deterministic business rules are often better for things that must happen exactly the same way every time, such as calculations, permissions or financial logic.
Good systems use the right tool for the job rather than forcing AI into every process.
Measure the result in time and reliability
An automation is valuable when it creates a measurable operational improvement.
That may be fewer admin hours, faster response times, fewer missed follow-ups, cleaner data or a more consistent customer experience.
The purpose is not to say the business uses automation.
The purpose is to make the business easier to operate.
Start with the boring repetitive work.
That is usually where the biggest practical gains are hiding.