What to automate in your business, and what to automate first.

What to automate first, and why the choice matters

Your first automation sets the tone. If it saves real time and works reliably, your team trusts the next one. If it targets something rare, messy or political, it stalls, and automation gets filed under "tried that, didn't work."

The goal of the first project is not to automate the most impressive thing. It is to automate something that happens often, follows clear rules and costs you every week. That is a less exciting brief, and it is the one that works.

Step 1: List your recurring tasks for two weeks

Memory is a poor guide to where time goes. The tasks that feel painful are often the rare, stressful ones. The real drain is usually the five-minute job done many times a week that nobody thinks about anymore.

For two weeks, ask everyone on the team (including you) to keep a simple running list. A shared spreadsheet works fine. For each recurring task, note:

• What the task is, in one sentence ("enter supplier invoices into QuickBooks")
• Roughly how long it takes each time
• How often it came up
• Which systems were involved: email, CRM, spreadsheet, accounting, a supplier portal
• What went wrong, if anything: a typo, a missed follow-up, a late reply

Two weeks is long enough to catch the weekly jobs and short enough that people keep it up. For monthly or quarterly work like month-end reporting, add it from memory and mark how often it happens.

Don't filter while you collect. "Copying addresses out of emails" belongs on the list as much as "building the quarterly report." Small tasks are where the surprises are.

Step 2: Score which business processes to automate

Once you have the list, score each task from 1 (low) to 5 (high) on five factors. You are not after precision. You are looking for the tasks that stand out.

Frequency

How often does it happen? Daily tasks beat monthly ones, because every minute saved repeats.

Time per instance

How long does it take each time, including switching between apps and checking the result afterwards?

Error cost

What happens when it goes wrong? A mistyped invoice amount, a lost lead or a missed deadline costs more than a misfiled photo.

Rule clarity

Could you write down how to do it on one page? If two people on your team would handle the same case the same way, the rules are clear. If the answer depends on who you ask, they aren't clear yet.

Data availability

Is the information the task needs already in a system, an email or a document? Or does it live in someone's head?

Add up the five scores and the highest totals become your shortlist. Then apply one sanity check: any task that scores 1 or 2 on rule clarity or data availability moves down the list, however high its total. A low score there means the process needs work before it can be automated.

If two tasks end up close, break the tie with two questions. Which one does your team complain about more? And which one touches customers? A task people are glad to hand over gets adopted faster, and a task that speeds up a customer's experience (a faster reply, a quicker invoice, a reminder that prevents a no-show) tends to show its value sooner than a purely internal one.

An example scoring table

The table below is an illustrative example for a fictional small service business. The tasks and scores are made up to show how the method works. Your own list will look different.

Look at what happens. Invoice entry and inquiry replies rise to the top: they're frequent, rule-based, and the data is already sitting in emails and PDFs. Discount decisions score high on error cost but low on rule clarity, so they stay with a person. The insurance paperwork is painful, but it happens once a year, so an automation would likely take longer to build and maintain than it would ever save.

Start boring

The best first automations are usually the dullest items on the list: data entry, reminders, filing, routing and status updates. That isn't a coincidence. Boring tasks tend to be frequent, rule-based and low-drama, which is exactly what makes them easy to automate well and easy to check.

Starting boring also builds the habits bigger projects need: clean data, clear rules, an exception path and someone who owns the result. Once invoice entry runs on its own, your team has seen how automation behaves, and the next project (perhaps an agent that drafts replies to customers) is a much smaller leap.

Good boring candidates include:

• Moving information from forms and emails into your CRM, and sending follow-ups on time (see CRM and follow-up automation)
• Reading invoices, receipts and intake forms into structured data (see document processing)
• Reminders that go out on a schedule
• Weekly reports assembled from several systems