How to find the first task to automate in your company

Once you start looking, you find plenty of work to automate. The hard part is choosing where to start. Pick a task that is too big, too rare or too vague, and the project drags on. Pick the right one, and it is live within days.

We use this test in our first working session with a client. You can also run it yourself with a list of what your team did last week.

Four signs a task is ready to automate

Go through your team’s week and note every task that passes these four checks:

• It repeats every week. The same job, again and again, not once a quarter.
• It follows rules you could write down. If you can explain it to a new hire as a list of steps, a system can follow the same steps.
• The information already lives in digital tools. Email, forms, spreadsheets, your CRM, your accounting software. If the work starts on paper or in someone’s head, that is a separate project.
• It costs real hours. A two-minute task done once a month is not worth the effort. A 10-minute task done 30 times a week deserves a closer look.

The tasks that pass are often the dull ones: copying details from one tool to another, renaming and filing documents, sending the same reminder, building the same weekly report. Dull is a good sign. Dull means predictable, and predictable work is what a system does best.

Put a number on the cost

Before you automate anything, measure it. You need four figures: how long the task takes, how often one person does it each week, how many people do it and how many weeks your team works in a year. We use 48 working weeks to leave room for vacation and statutory holidays.

The formula fits on one line:

Minutes per task × times per week × people × 48 working weeks ÷ 60 = hours per year

Say a team of three copies details from incoming emails into a spreadsheet. Each person does it 20 times a week, and it takes 15 minutes each time. That is 15 × 20 × 3 × 48 ÷ 60 = 720 hours a year, or 96 working days of 7.5 hours. At 37.5 hours a week, a full-time year is 1,800 hours, so this one task takes up 0.4 of a full-time job.

Ask the people who do the task to time it for a week instead of guessing. The goal is not precision. The number lets you rank tasks against each other and decide whether a fix is worth what it costs.

What to leave alone

Three kinds of work should stay with people, at least at first:

• Judgment calls. Pricing an unusual request, deciding whether to take on a client, handling a complaint. These depend on context that rarely lives in a tool.
• Rare exceptions. If a case comes up twice a year, handle it by hand. Building for every odd case turns a simple system into a fragile one.
• Relationship work. Calls with clients, hard conversations, the note that has to come from you. This is the work your team should have more time for, not less.

Automation can still help around these tasks. It can gather the facts before a decision, or send an exception to the right person. It should not make the call.

What a good first win looks like

A good first automation is small, visible and easy to check. Small, so it is live in days, not months. Visible, so the people who did the work by hand notice the change in their week. Easy to check, so anyone can compare the result with what a person would have done.

It runs inside the tools your team already uses, so nobody has to learn new software. We test it on real cases from your last few weeks before it goes live. It comes with a short guide in plain language. And it has an off switch: if something looks wrong, anyone on the team can turn it off and go back to doing the task by hand.

After a few weeks of running well, you have more than hours back. You have a team that trusts the system and a short list of what to automate next. That trust makes the second project easier than the first.

If you already have a task in mind, send it to us. Within 24 hours, we reply with suggestions on how to tackle it and tell you plainly whether it is worth automating. If you want to go further, we can talk it through on a call.