What to automate first in a small business

Automate the task your team does often, by the same rules every time, that people quietly dread. Not the biggest problem or the most exciting idea, but the boring, predictable job that eats a little time every day.

That is usually where a first automation pays off fastest and causes the least trouble. Here is how to find it, score it and pick one.

What to automate first in a small business

A good first candidate has four traits.

Repetitive. The same steps happen again and again.
Rule-based. You could write the steps on one page, and two people would follow them the same way.
Frequent. It happens daily or weekly, not a few times a year.
Annoying. People put it off, rush it or complain about it.

If a task has all four, it is worth a close look. If it is not rule-based, it is not a good first project. Automation is most reliable, and easiest to check, when the steps are clear.

Common first wins

These are common places to start:

Copying data between systems. Typing the same customer details into your customer list, your accounting software and a spreadsheet.
Reminders and follow-ups. Appointment reminders, overdue invoice nudges and "just checking in" emails after a quote.
Intake forms. A web form or email that someone reads, then retypes into another tool.
Status updates. Customers or staff asking if an order shipped or a job is booked, when the answer already lives in a system.
Reports put together by hand. Pulling numbers from three places every Monday and pasting them into one sheet.

Each one follows clear rules and is easy to check, as long as someone looks at the results for the first while.

What not to automate first

Some work looks tempting but makes a poor starting point.

Judgement calls. Pricing a custom job, deciding which customer gets priority, handling a complaint. Software can help gather the facts, but a person should make the call.
Messy processes nobody agrees on. Automating a broken process just makes the mess happen faster. Agree on the steps with your team first.
Rare tasks. Year-end jobs and one-off projects do not come around often enough to be worth the setup and upkeep.
Anything costly or hard to undo if it goes wrong. Charging customers, cancelling orders or changing prices. Save these for later, once you trust the approach.

None of these are off the table forever. Your first project should teach your team how automation fits your business without putting anything important at risk.

How to score your candidates

Sit down with two or three people who actually do the work. List every task that feels like a candidate. Then score each one from 1 (low) to 3 (high) on four questions:

How often? Daily scores high. Monthly scores low.
How long? A few minutes scores low. A big part of someone's day scores high.
How error-prone? Rare mistakes score low. Regular typos, missed steps or forgotten follow-ups score high.
How painful? If nobody minds it, score low. If everyone puts it off, score high.

Add up the scores. The top of the list is your shortlist.

Before you pick, check one more thing: does everyone agree on how the task is done today? If three people describe it three different ways, it is not ready yet. Choose the highest-scoring task where the steps are already clear.

You are not looking for a perfect score. You are looking for a clear, contained task you can finish, test and measure.

Example (hypothetical): a small Alberta trades company

Imagine a heating and plumbing company in central Alberta with about a dozen staff. The owner wants to "automate the office," which is too big to act on.

The office manager and two technicians list their candidates:

Retyping online booking requests into the scheduling tool
Sending appointment reminders the day before
Chasing unpaid invoices
Building a weekly jobs report for the owner
Quoting larger renovation jobs

Quoting scores high on pain, but every quote is a judgement call, so it comes off the list. The weekly report is annoying, but it scores lower on how often than a daily task.

Retyping booking requests scores high on all four questions. It happens every day, takes real time, leads to typos in addresses and phone numbers, and nobody likes doing it.

So they pick one thing. New web bookings go straight into the scheduling tool, and the office manager checks each one instead of retyping it. Before the change, she writes down roughly how long the retyping takes in a normal week, so they can compare later.

Once that is running and trusted, appointment reminders are next on the list.

Start small, measure, then do the next one

Your first automation is a test of the approach, not a big overhaul. Keep it small enough to finish.

Measure before and after. Note how long the task takes now, then check again a few weeks after the change.
Keep a person in the loop at first. Have someone review the output until it earns their trust.
Write down what it does. One page is enough, so the next person knows how it works.
Then pick the next task. Go back to your scored list and repeat.

This is how we think about it at LMG Labs: one useful thing, then the next. If you are weighing whether to use the tools you already pay for or build something new, our post on off-the-shelf vs custom software walks through that choice.

What to do this week

Ask your team which task they would most like to never do again.
List every candidate you can think of, then score each one on how often, how long, how error-prone and how painful.
Cross off judgement calls, rare tasks, anything nobody agrees on, and anything costly or hard to undo.
Pick the top task and write down how long it takes today.
Check whether the software you already use has a built-in option for it.

If you want a second set of eyes on your shortlist, tell us what you are working on.