What Should You Automate First? A Worksheet for Small Business Owners
You finish the job, then spend four minutes copying the customer's details into another system. Tomorrow you'll do it again. Meanwhile, someone has suggested an AI receptionist, a new CRM, and a chatbot for your website.
Start with the copying.
After 30 years working with technology, my advice is to choose a task you can describe clearly and test in a small way. You don't need to redesign the whole business to find out whether automation helps. You need one job, a starting measurement, and someone responsible for checking the result.
What makes a useful first automation?
A useful first candidate repeats often, follows clear rules, and has an outcome you can check. Give extra weight to tasks that are easy to pause or handle manually if the first attempt needs adjusting.
Copying approved information between two tools might qualify. Deciding how to handle an unhappy customer probably needs a person, even if software gathers the details beforehand.
But frequency alone isn't enough. A daily task with unclear rules can create daily confusion when automated. If you're still weighing the broader decision, start with these questions about whether automation fits your business.
How do you compare the tasks on your list?
Write down four recurring tasks and answer the same questions for each. Use actual work from this week, including the interruptions and corrections that don't appear on a job description.
Copy this worksheet into a document or notebook. Fill one out for each task.
| Question | Your answer | |---|---| | What exact task are we doing? | ____________________ | | How many times per week? | ____________________ | | How many minutes each time? | ____________________ | | What happens if it's missed or wrong? | ____________________ | | Which tools or records are involved? | ____________________ | | Which steps need a person's judgment? | ____________________ | | Who will check the first test? | ____________________ |
Count the small jobs too. For a Vermont service business, that could mean copying a confirmed appointment, chasing an unanswered estimate, or assembling a weekly report from two spreadsheets.
Don't write "fix scheduling." Write "copy confirmed appointments from the booking tool into the crew calendar." That gives you something concrete to investigate.
How much time is actually tied up in it?
Multiply the weekly count by the minutes per occurrence. The result is time spent today, not a promise of time saved.
Here's a hypothetical example. A shop copies customer details 12 times a week, taking four minutes each time:
12 x 4 minutes = 48 minutes per week spent on copying.
Some of those minutes may still be needed for checking, exceptions, or corrections after automation. Subtract that remaining work when you measure the result. Setup and maintenance take time too.
And if your estimate is a guess, mark it as a guess. Time the next few occurrences before spending money. A rough measurement you actually took is more useful than a polished savings claim you can't explain.
Which task should come first?
Choose a repeated task with a clear starting event, clear rules, and a manageable consequence if the test fails. It doesn't have to be the biggest number on the page.
Use these four questions to narrow it down:
- Can someone explain the steps without saying "it depends" at every turn?
- Do we have the information needed to do it correctly?
- Can we check the output before it affects a customer or the books?
- Can we pause it and return to the manual process?
If you're answering no, simplify the task first. That's a process decision. Buying another tool won't make the missing rule appear.
Our guide to common small business automation mistakes covers what happens when a build starts too broadly.
What should the first test look like?
Write one sentence that describes the test: "When this happens, prepare that, and let this person check it."
For the hypothetical copying task, the first version could prepare a record for staff review. Check that required fields are present, duplicates are avoided, and a correction has somewhere to go. Keep a short log of what worked and what needed fixing.
One task. One owner.
If the tools can't exchange the information reliably, that's a useful finding. Connecting your existing tools starts with checking what they support. If the work instead depends on unclear customer messaging, Get AI-Ready addresses customer personas, brand voice, and positioning.
No. A fixed rule or an existing software feature may be enough. Choose the simplest approach that performs the task reliably and lets your team check it.
Not automatically. A smaller task with clear rules may make a better first test. Consider the time involved alongside the judgment required and the consequences of an error.
Start with someone reviewing the proposed message before it goes out. Decide what should stop the sequence, such as a reply or a changed appointment, before sending automatically.
We bill for our time and materials, and we'll talk through the rate once we know the job. Check the likely benefit and scope first. A small task may be something your team can fix using software you already have.
Bring your completed worksheet when you start a conversation with us in Essex Junction. We'll look at one task and help you decide what to try. When you're ready to compare the work involved, here's how we bill.
More from the blog
Want to know what this would take for you?
Tell us the task and the tools you use. We’ll look at how your business runs and tell you what we’d build.
Start a conversation