Most businesses do not stall on automation because the technology is hard. They stall on the first decision: out of everything that could be automated, what should come first? Pick wrong and the project drags, the team loses faith, and the whole idea gets quietly shelved. Pick well and you get an early, visible win that funds everything after it. So the goal for your first automation is not the most impressive thing. It is the thing most likely to work.
Here is a one-page way to make that choice. It fits on a napkin, and it works for a two-person shop or a two-hundred-person company.
Score each candidate on three things
List the tasks you are tempted to automate. For each one, rate three factors from low to high. You do not need precise numbers. A quick one-to-five gut rating is enough to separate the strong candidates from the weak ones.
- Time saved per week. How many hours does this task swallow across everyone who touches it? A task that eats five hours a week is worth far more than one that eats twenty minutes.
- Frequency. How often does it happen? Something that runs many times a day compounds. A once-a-quarter task, even a painful one, has fewer chances to pay you back.
- Drain. How much does the task wear on the people doing it? Repetitive, error-prone, morale-sapping work carries a hidden cost beyond the clock. Removing it buys you goodwill and fewer mistakes, not just hours.
Multiply those three together in your head. High time, high frequency, high drain is your ideal target. That is a task begging to be handed off.
Weigh it against complexity
Now add the counterweight: how hard is this to actually automate? A task with clean, consistent steps and data already living in one place is simple. A task tangled across five systems, full of exceptions and human judgement, is complex.
Your first automation wants to sit in the sweet spot: high value, low complexity. Strong payoff, quick to ship, unlikely to stall. Save the ambitious, tangled work for later, once you have proof and momentum behind you.
Your first automation is not the one that impresses people in a meeting. It is the one that quietly works by Friday.
Two traps to avoid
The first trap is the exciting complex project. There is always a grand idea that would transform the business if it worked. It is tempting because it is the reason you got interested in automation at all. But it is slow, risky, and the easiest thing to stall on. Starting there means months before anyone sees a result, and plenty of chances to lose confidence along the way.
The second trap is automating something so rare it never pays back. A task might be genuinely annoying, but if it only happens a few times a year, the effort to automate it will outlast the time it saves. Frequency is what turns saved minutes into real hours.
A worked example
Say you are weighing three tasks. Manually chasing overdue invoices happens weekly, eats a few hours, and everyone hates doing it: high time, high frequency, high drain, and the steps are fairly clean. That scores well and is simple to build. Automating your annual budget model saves a big chunk of time, but only once a year and it is genuinely complex: high value per run, but low frequency and high complexity, so it waits. Re-typing new leads from a form into your CRM happens many times a day, takes a couple of minutes each, drains focus, and is trivially simple: modest time per instance, but frequency and simplicity make it a clear early win.
Ranked honestly, invoice chasing and lead entry rise to the top. The budget model, however appealing, drops to the bottom of the list. The framework just told you where to start without a single meeting.
Start where you can win
The point of a first automation is momentum. Choose the task that is valuable enough to matter, frequent enough to pay back, and simple enough to ship without drama. It builds confidence across the team and frees up the very time you need to tackle the bigger, harder projects next. One clean win funds the next one. That is how automation actually takes hold in a business, one deliberate choice at a time.