What to automate first when it's you and two other people

Almost every time someone tells me "I want to automate my business," they already have a tool in mind. A chatbot. A CRM. Some system a cousin recommended.

And almost every time, they're starting at the end.

Automating isn't buying software. It's taking off your plate a task that repeats, doesn't require judgment, and that you're currently doing yourself. The question isn't which tool, it's which task. And in a three-person business, not all tasks are worth the same.

The three-question test

Before automating anything, take a sheet of paper and write down everything you did yesterday. All of it. Answering WhatsApp, logging sales, going to the bank, sending the order to your supplier, collecting payment, closing out the register.

Now cross out anything that doesn't pass all three:

1. Does it repeat? If it happens once a month, don't automate it. Put it on the calendar and move on. 2. Is it done the same way every time? If you decide differently each time, it isn't automatable yet. First it becomes a process, then it gets automated. Automating a mess just gives you a faster mess. 3. Are you the one doing it? If someone else does it and does it well, it isn't the urgent one. The urgent one is the task that has you working as a data-entry clerk.

What survives all three is your real list. Usually it's two or three things, not twenty.

The order I follow

When the team is two or three people, the order is almost always this.

### First: recording what you sell

Not because it's the most annoying, but because it's the one costing you decisions. If your sales live in a notebook, you can't know which product carries the best margin, which day is the good one, or whether this month beat the last. You're driving at night with the headlights off.

And note: automating the record doesn't mean buying an expensive point-of-sale system. In many cases it starts with a well-built shared spreadsheet, the right columns and one formula. What matters isn't the tool — it's that the data gets captured once, at the moment the sale happens, and not at night from memory.

### Second: the answer you give fifty times a day

Price, hours, location, whether you deliver, whether you take cards. If you're typing the same answer every day, that should no longer be coming out of your fingers.

This is where WhatsApp Business earns its keep: a greeting message, an away message with your real hours, and quick replies with shortcuts. It takes an afternoon to set up and gives you back hours every week. It isn't glamorous and you won't be showing it off to anyone, but early on it's what frees the most time.

An AI chatbot can come later. Before that, the answers have to be written down — and that exercise alone solves half the problem.

### Third: the follow-up you keep forgetting

The customer who asked and didn't buy. The quote you sent that nobody answered. The regular who stopped coming.

That money was practically in your hand and it walked because you forgot. Automating here can be as simple as a "next contact" column in your sheet and a calendar reminder. You don't need a three-hundred-dollar-a-month CRM to remember three people.

### Fourth: getting paid

Invoices, payment reminders, payment links. Automating collections almost always pays for the project on its own, because it shortens the days between the sale and the money landing in your account. I put it fourth rather than first because it depends on the other three being in order: you can't automate collections if you aren't clear on what you sold and to whom.

What is almost never worth automating first

• ·Social media. Scheduling posts feels productive and rarely moves the needle for a local business. It comes later, not first.
• ·Pretty reports. A dashboard full of charts is worthless if the underlying data is captured badly. Data first.
• ·Anything only you know how to do. If a task depends on your judgment, automating it is an expensive mistake. That one gets documented and delegated, which is not the same thing.

One rule so you don't get this wrong

If a task takes less than an hour a week, don't automate it yet. The time you'll spend setting it up, learning it and fixing it is more than the time it saves.

Start with the task costing you four hours a week or more. That one's worth it. There the math works in your favor without you having to talk yourself into it.

How we handle this in a diagnostic

In the 45 minutes of the diagnostic we map exactly this: which tasks repeat, how much time they take today, and which ones pass all three questions. Out of that comes a list ranked by impact against effort, and we decide what gets built first.

Nothing gets implemented before we've written down how much time that task takes today. That's the baseline. Without it there's no way to show afterwards that the automation worked — and that proof is exactly what makes it worth doing.

Want to know what you'd automate first? The diagnostic is 45 minutes, it's free, and you leave with the ranked list whether or not you work with us.