Is AI automation worth your time?
Before you spend evenings (or money) on automation, it helps to do a bit of napkin maths. Not spreadsheet theatre. Just enough to see whether the hours you might get back are in the same ballpark as the cost.
Why bother doing the maths?
Automation sounds clever until you realise you've paid for a tool that saves twelve minutes a fortnight. Putting rough numbers on it keeps you honest. If the time you get back wouldn't cover the setup and the monthly AI bill, you can spend that energy somewhere else.
This calculator is deliberately rough. It does not know your software, your staff, or the weird edge cases in your week. Treat it as a gut check, not a quote.
What it doesn't capture
Hours and dollars are the easy part. A good automation also means fewer mistakes, less stress on a Sunday night, and jobs that used to sit in the "never quite get to it" pile. Those things matter, and they do not show up cleanly in a calculator.
So if the numbers come out close, use your judgement. If they come out clearly bad, believe them.
When it's not worth automating
- The task barely happens, or only a few times a year.
- Every job is different and needs a human call each time.
- Getting it wrong costs more than doing it slowly by hand.
- You would still spend ages checking every output, so you have not really saved the time.
- The tool cost eats the saving, and there is no other benefit that makes up for it.
If that sounds like you, skip the build. I would rather tell you that on a call than spend a day on something that does not pay off.
A worked example (coaching business)
An anonymised coaching business books about 45 to 50 courts a week, mostly in a one-hour weekly booking window. About 20% of those bookings (9 or 10 a week) get moved and need rebooking or cancelling, and group training each week means cancelling a court and shuffling coaches.
Reading bookings and the credit balance already works. Automated rebooking and cancelling is still in progress, so the time figures below are projected, not measured after go-live.
- About 2.5 hours a week on rebooking and cancelling (modelled as 10 moves × 15 minutes).
- Roughly 95% of that work looks automatable once booking and cancelling are built.
- About 10 to 15 minutes a week left for someone to log in and approve changes (the example uses 12).
- AI tool cost about $30 AUD a month in the example (Claude Pro). Your business cost may be higher depending on enterprise plans and data sensitivity.
- One-off setup left at $0 (the free day). Put in a figure if you are costing a paid build or your own time.
- Hourly cost set at $40 AUD as a stand-in for whoever does the admin (include the owner's time).
Hit Load coaching example in the calculator to drop those figures in, then change anything that does not match your business.
Rough numbers calculator
This is a rough-and-ready estimate, not a precise forecast. Tweak the figures to match your week. All numbers stay in your browser.
You'd save about 2.3 hours a week and break even in about less than a week.
- Hours of admin per week
- 3.7 hours
- Hours saved per week
- 2.3 hours
- Hours saved per year
- 120.5 hours
- Saved per week (before AI cost)
- $93
- Saved per month (before AI cost)
- $402
- Saved per year (before AI cost)
- $4,819
- Saved per week (after AI cost)
- $86
- Saved per month (after AI cost)
- $372
- Saved per year (after AI cost)
- $4,459
- Break-even on setup
- About less than a week
- Net position after 12 months
- $4,459
Want a hand finding the right task?
If the rough numbers look promising, have a look at working with me, book a time, or email me at anaru.dev@gmail.com. No hard sell. If it is not worth automating, I will say so.