How do I know if a task is worth automating?

Rate it 1 to 5 on volume, repetitiveness and how rules-based it is, then multiply the three. High combined scores are worth automating; a task that is rare, always different, or full of judgement calls usually isn't.

Should I automate my most annoying task first?

Not necessarily. The most annoying task is often annoying because it needs judgement, which makes it hard to automate well. Start with the highest-scoring boring, repetitive, rule-based task instead — it's the safest early win.

Claude Thinks → Framework

What to automate first in your business

Q: What should a small business automate first?

Automate the task that scores highest on three factors together: how often it happens (volume), how similar it is each time (repetitiveness), and how clear the rules are (rules-based). A high-volume, near-identical, rule-following task is the ideal first candidate.

A simple scoring framework · Takes 10 minutes to run

Short answer: Automate whichever task scores highest across three things at once — how often it happens (volume), how similar it is each time (repetitiveness), and how clear the rules are (rules-based). A task that's high on all three is your safest, most valuable first automation. Score your tasks 1–5 on each, multiply, and start at the top of the list.

Most owners automate the wrong thing first — usually their most annoying task. But annoying often means "needs my judgement", which is exactly what's hard to hand off. The right first pick is boring, frequent and predictable. Here's a simple way to find it.

The formula

Priority = Volume × Repetitiveness × Rules-based

Score each task from 1 (low) to 5 (high) on all three, then multiply. Because it's a multiplication, a task has to score decently on all three to rise to the top — a 5 for volume means nothing if it's a 1 for rules. That's the point: it filters out the ones that look tempting but will fight you.

The three factors

1. Volume — how often does it happen?

Ten times a day scores a 5. Once a quarter scores a 1. Automation earns its keep through repetition, so frequency is the biggest lever on payoff.

2. Repetitiveness — how similar is it each time?

Nearly identical every time (same steps, same shape of input) scores a 5. Different every single time scores a 1. The more it varies, the harder it is to automate reliably.

3. Rules-based — how clear are the rules?

"If X, always do Y" with no judgement scores a 5. "It depends on the customer, the mood, the context" scores a 1. Clear rules are what a machine can follow safely; fuzzy judgement is what keeps you in the loop.

Worked example: scoring a few common tasks

Here's the framework applied to typical small-business tasks (your numbers will differ — that's the whole idea):

Task	Volume	Repetitive	Rules-based	Score
Sending booking confirmations	5	5	5	125
Chasing overdue invoices	4	4	4	64
First-draft replies to FAQs	5	4	3	60
Writing custom proposals	3	2	2	12
Handling a customer complaint	2	1	1	2

The complaint scores lowest despite being the most stressful — because it needs judgement every time. The confirmations win: dull, constant, and totally rule-based. That's your first automation.

Quick rule of thumb: if a task is boring you because it's the same thing over and over, it's a great candidate. If it's stressing you because every case is different, keep a human on it — for now, that human is you.

How to run it yourself

List every recurring task that eats your week — brain-dump, don't filter.
Score each 1–5 on volume, repetitiveness and rules-based.
Multiply the three numbers for each task.
Sort highest to lowest. The top one or two are your starting line.
Prove it manually with AI first (see how to automate your admin), then decide whether it's worth wiring up hands-off.

A caution before you automate

A high score means a task is a good candidate — not that automating it is risk-free. Before you make anything run on its own, ask: what happens if it gets one wrong? For a booking confirmation, not much. For something touching money, contracts or safety, the bar is much higher and a human check stays in the loop. High volume plus high stakes is exactly where it pays to have it built properly rather than rigged up quickly.

Want this set up and running — not just scored on a page?

Once you know your top task, the next step is building it so it runs reliably and safely. That's what SG1 Consulting does — taking your highest-scoring task from "should automate" to "quietly handled".

Talk to SG1 →