Free workflow

Find the right first automation — before you build the wrong one.

A five-step audit for finding and building your first AI-powered process that actually holds up. The right first automation is boring. That's how you know it'll work.

The problem

The first automation most people build is the wrong one. They pick something exciting, something visible, something that sounds impressive when they describe it to their team. And it breaks within a week — because exciting processes are exciting precisely because they require judgment, context, and the kind of nuance that no system handles well on its first pass.

The right first automation is boring. Data entry, formatting, file routing, status updates. Boring means structured, repetitive, and low-stakes enough that you can learn how your system fails without anything catching fire.

The audit

Track what you actually do — 3 days of honest logging, not what you think you do
Score on three dimensions — volume, judgment, structure
Build a first pass — ugly, minimal, handles the core loop only
Test with real data — yesterday's work, not sample data
Measure honestly — the 3:1 rule for automation viability

Get the full system.

The complete audit with scoring templates and worked examples. Enter your email and the system reveals below.

01 — Track what you actually do

3 days. Honest logging.

Not what you think you do. What you actually do. Set a timer, log every task, note how long it takes.

Most people discover that 30–40% of their day is stuff they'd never consciously chosen to spend time on. This is your automation surface area — and you can't see it without tracking.

02 — Score on three dimensions

Volume. Judgment. Structure.

Volume is how often you do it. Judgment is how much thinking it requires each time. Structure is how consistent the inputs and outputs are.

High volume, low judgment, high structure is your target. That's the sweet spot where automation pays for itself immediately. Score each task 1–5 on all three dimensions.

03 — Build a first pass

Ugly is correct.

Not the final system. A draft. Something ugly that handles the core loop and nothing else.

Resist the urge to make it elegant. Elegance comes from iteration, not from planning. The first pass exists to reveal your assumptions — and assumptions are where systems break.

04 — Test with real data

Yesterday's work, not sample data.

Not sample data, not test cases. Actual work from yesterday. This is where most automations reveal their assumptions — the edge cases that seemed unlikely until they showed up in production.

05 — Measure honestly

The 3:1 rule.

If the automation doesn't save at least three times the effort it takes to maintain, it's not ready. Plenty of automations pass the “cool demo” test and fail the 3:1 test. Kill them early.

Honest measurement is the difference between automation that compounds and automation that becomes another thing to maintain.

Get the full system

Install in 30 seconds.

The complete audit — scoring templates, worked examples, the 3:1 measurement framework, and decision criteria for when to automate vs. when to leave it alone — is a Claude Code skill you can install and run on any process.

Install from GitHub

Want this as part of a broader system?

The Automation Audit is one piece. I build AI-native systems for operators — research, prospect intelligence, content pipelines, signal detection — each with the same structural guarantees. If you want help implementing this for your team, let's talk.

hello@thinkmetis.co