How to Spot What to Automate First in Your Business

Most business owners already know something in their business should be automated. What they can't work out is which thing. So the project stalls, the spreadsheet keeps getting rebuilt every Friday, and another year goes by.

This is the process I walk through with every business before I build a single thing. You can do it yourself in an afternoon with a notepad.

Start with the calendar, not the wish list

The instinct is to start with what annoys you. That's a bad filter — annoying and expensive are different problems.

Instead, take one week and note every task in your business that repeats. Not projects. Repeats. Anything someone does daily, weekly or every single time a customer comes through the door.

For each one, write down three numbers:

• How often it happens (per week or per month)
• How long it takes each time
• Who does it, and roughly what that hour costs you

A 6-minute task done 40 times a week is four hours. That's half a day, every week, forever. Meanwhile the two-hour job that happens once a quarter feels far worse and costs you almost nothing. Frequency beats irritation nearly every time.

The four questions that identify a good candidate

Once you've got your list, run each task through these four questions. A task that gets four yeses is where you start.

1. Does it follow the same steps every time?

If a task has clear rules — take this information, put it there, send it to them, file it here — it can be automated. If it needs judgement, negotiation or a relationship, it can't (and shouldn't be).

2. Is someone rekeying information that already exists?

This is the single biggest source of wasted hours in small business. A customer fills in a form, then someone copies those details into the CRM, then into a document, then into an invoice, then into a spreadsheet. Same data, four times, four chances to get it wrong.

Anywhere information gets typed twice is an automation opportunity with a guaranteed return.

3. Does a delay here cost you money?

Not all slow tasks are equal. A late internal report is a nuisance. A lead enquiry sitting unanswered for six hours is revenue walking out the door — the business that replies first usually wins the job.

Tasks that sit between a customer and a decision get moved to the top of the list.

4. Would it still need doing at 2am, on a public holiday, or when your best person is on leave?

If the answer is yes, that task shouldn't depend on a human remembering. Reminders, follow-ups, confirmations, chasing paperwork, renewal dates — these are the jobs that quietly fall over the moment someone's away.

The shortlist test

Score each task out of four. Then look at your top few and ask one last question: if this ran by itself starting Monday, what would change?

If the honest answer is "not much," park it. If the answer is "we'd get three more hours a week and stop dropping enquiries," you've found your first build.

Almost every business I've worked with lands on something in one of these three buckets:

• Enquiry and lead handling — responding, qualifying, logging, following up
• Document and data admin — forms, letters, quotes, reports, stock lists, anything rebuilt by hand
• Reminders and repetitive chasing — the recurring nudges nobody enjoys and everybody forgets

A logistics business I worked with was hand-building the same operational report every week. An automotive workshop was chasing bookings and follow-ups manually. A real estate agency was rekeying the same client details across three separate documents before anything could be signed. Different industries, same shape of problem: information that already existed, being moved by hand.

What not to automate first

A few things are worth deliberately leaving alone:

• A broken process. Automating a bad workflow just makes it fail faster. Fix the process, then automate it. This is why I look at how something actually runs before I build anything.
• Anything that's genuinely someone's judgement call. Pricing a difficult job, handling an unhappy customer, deciding whether to take on a client. Automation should clear the desk so your people can do more of this, not less.
• The thing you do twice a year. It'll never pay for itself, no matter how much you dislike it.

Where AI actually fits

Worth being direct about this, because it's the question I get asked most.

For most of the work I do, the client never touches AI. What they get is a workflow that runs on its own — the form fills the document, the document goes for signature, the CRM updates, the reminder fires. It's their existing tools, connected properly, doing the boring parts without anyone driving.

AI is a tool I use behind the scenes to design and build those workflows faster and cheaper than the traditional approach. Where AI genuinely helps the client directly — a bot that handles inbound enquiries around the clock, for example — I'll use it and say so. Where it doesn't, I won't bolt it on for the sake of it.

You're buying hours back and fewer things falling through the cracks. How it's built is my problem, not yours.

Start small, prove it, then expand

The best first automation is boring, specific and measurable. One workflow. Something you can point at in a fortnight and say "that used to take four hours a week and now it takes none."

That's the whole argument. Once one thing runs by itself and nothing breaks, the next three are easy decisions.

If you'd like a second set of eyes on your list, that's essentially what a free automation audit is — I look at how your business actually runs, tell you which one or two things are worth automating first, and tell you honestly if the answer is "none of it yet."

Try the free tool

I've turned the shortlist method in this article into a two-minute interactive tool: the Automation Priority Scorer. Add your recurring tasks, answer the four questions for each, and it ranks them into a shortlist with the hours each one is costing you. No email, no sign-up — your list stays in your browser.