Worth It or Not? The Automation Decision
Turn your Top 5 into verdicts: the worth-it math, the automate/delegate/eliminate framework, and a decision matrix that ends the guessing.
In this lesson, you'll learn to…
- 1Run the worth-it math: time saved versus setup time — the honest ROI of any automation
- 2Apply Automate / Delegate / Eliminate: the framework that catches what automation-thinking misses
- 3Place any task on the simple decision matrix: frequency × judgment
- 4Choose between build-with-AI, buy-a-tool, and ignore-for-now for each candidate
- 5Finish Module 1 with verdicts on your Top 5 — and know exactly where the academy takes you next
The question that saves you from yourself
"Is automating this actually worth the setup effort?"
What a mature question — and asking it now, BEFORE building anything, puts you ahead of most people who ever touch automation. Because here's the failure mode nobody warns beginners about: it isn't automating too little. It's enthusiasm-driven automating — spending five hours setting up a workflow that saves ten minutes a month, burning out on the setup, and concluding 'automation isn't for me.' The tool didn't fail; the decision did. Nobody ran the math.
You're arriving at this question with everything it needs: a marked Work Log (what AI can realistically take — Lesson 2) and a costed Top 5 with monthly hours attached (Lesson 3). Today adds the final piece of Module 1's promise — *look at your own week and point at exactly what a machine should be doing* — by teaching the decision itself: the honest math, a three-option framework that's bigger than automation, and a simple matrix that turns 'should I?' into a square on a grid.
One reassurance before the arithmetic: 'not worth it' is a successful outcome of this lesson. A considered no — with reasons — protects your energy for the yeses that pay. The goal was never to automate everything. It was to automate the RIGHT things.
Key idea: The beginner trap isn't automating too little — it's enthusiasm-driven automating: five hours of setup for ten minutes of monthly savings, then burnout. The decision, not the tool, is what fails. So we learn the decision.
The worth-it math
The core arithmetic is one honest comparison: what it saves versus what it costs.
What it saves: your column D number — monthly hours — projected forward. A task saving 3 hours a month saves 36 hours a year; automation payback compounds because the task was going to repeat forever.
What it costs: three things, and beginners always undercount the last two. *Setup time* — learning, building, testing; be pessimistic, then add half. *Ongoing attention* — no automation is fully free: reviewing output, occasional fixes, keeping instructions current. A rough rule for beginner-sized automations: budget ~10–20% of the saved time as permanent review-and-maintenance attention. *The mistake cost* — what's the worst realistic error, and what would it cost? A typo in an internal note costs nothing; a wrong number sent to a customer costs real money and trust. High mistake-cost doesn't forbid automation — it demands the human-review kind (AI drafts, you decide), which raises the ongoing-attention cost. Price that honestly.
The verdict rule of thumb: if the setup pays for itself within about two months of saved time — and the mistake cost is survivable with review in place — it's worth it. Devon's descriptions: 2.7 hours saved monthly, maybe 2 hours of setup, low stakes with his final read — pays back in three weeks; obvious yes. His dramatic brand negotiation: 2 hours a month, sky-high judgment, relationship-critical — the math never starts, because Lesson 2 already ruled it. The math is quick. The honesty is the skill.
Watch out: Two costs beginners forget: ongoing attention (~10–20% of saved time, forever — review is part of automation, not extra) and the mistake cost (price the worst realistic error, not the average day). Count both and the math stops lying to you.
Bigger than automation: Automate / Delegate / Eliminate
Now widen the lens, because 'should I automate this?' quietly assumes the answer to a better question: *should this task be handled differently at all — and by what?* Three options, always considered in this order:
Eliminate — always ask first. Some repeated tasks exist out of habit, not need. The report nobody reads. The status meeting that a message could replace. The double-entry that started as a workaround years ago. Before automating anything, ask the blunt question: *if this task quietly stopped, who would notice — and would it matter?* If the honest answer is 'nobody, and no,' you just saved 100% of the time with zero setup. Automating a pointless task doesn't remove the waste — it gives the waste a machine.
Delegate — the middle option automation-thinking forgets. Some tasks need a human, just not YOU: they need hands and care, but not your specific judgment or context. If you have (or could have) help — a VA, a teammate, a family member in the business — a written-down task with clear steps delegates beautifully. And notice: the seeing skills you've built this module are exactly what make delegation work too. A task you can explain is a task you can hand off — to a person OR a machine.
Automate — for what remains. Repeats that pass the elimination question, don't fit a human helper, land in Lesson 2's families, and win the worth-it math. That's the pile with your name on it — and it's what Modules 2 through 6 exist to teach you to build, safely, step by step.
The A/D/E sequence
Pro tip: Ask in order: Eliminate first (zero setup, total savings), Delegate second (human hands, not your hours), Automate third (the machine takes what remains). Automation is the third question, not the first — which is exactly what makes it powerful when it wins.
The decision matrix: two questions, four squares
For the automate pile, one last tool makes the verdict almost mechanical: a two-question matrix. Question 1 — how OFTEN? (High: weekly or more · Low: monthly or less.) Question 2 — how much JUDGMENT? (Low: mechanical, 'almost the same' · High: decisions, relationships, your name — the never-three.) Every task lands in one of four squares:
| LOW judgment | HIGH judgment | |
|---|---|---|
| HIGH frequency | AUTOMATE NOW — the sweet spot: your first builds live here | AI-ASSIST — machine drafts the middle, you keep every decision (most real tasks land here) |
| LOW frequency | CHECKLIST IT — setup rarely pays back; a written checklist captures the win | KEEP IT MANUAL — rare + judgment-heavy is exactly what your freed hours are FOR |
Weekly-or-more AND mechanical — 'almost the same' every time
The sweet spot — Module 2 starts here
Frequent, but decisions or relationships woven through it
AI drafts the middle · you keep the judgment — the doctrine, applied
Rare — monthly or less, whatever the judgment level
Setup rarely pays back — capture it in a checklist, keep judgment work yours
Key idea: Your first automations — the ones Module 2 will teach you to build — come from the top-left square: high frequency, low judgment. Start where winning is easy; earn your way toward the harder squares as your skills and review habits grow.
Build, buy, or ignore — and Devon's five verdicts
One practical fork remains for each 'automate' or 'AI-assist' verdict: how will it get done? Three honest routes. Build-with-AI — for family work (drafting, extracting, transforming, answering), a general AI assistant plus your instructions handles it: no purchase, no code; this is what Module 2 teaches, and it's the right default for nearly everything on a beginner's list. Buy-a-tool — when the need is a specialized craft or heavy plumbing between apps; fine, but never first: a purchase before you understand your own workflow buys shelf-ware (Module 2 has a whole lesson on choosing without drowning). Ignore-for-now — a real decision, written down with its reason: the math failed, or the timing's wrong. Written-down ignores don't nag; vague maybes do.
Watch Devon's Top 5 get its verdicts — the whole module converging on one worksheet:
Video descriptions — AUTOMATE NOW (build-with-AI)
8×/month, mechanical, low stakes with his final read. High-frequency + low-judgment square; pays back in weeks. His Module 2 starter project.
Clip-hunting — AUTOMATE NOW (build-with-AI)
His biggest hours (5.3/month). AI reads the script and suggests moments; he makes the cuts. Extracting family, sweet-spot square.
Cross-posting — AI-ASSIST now, tool later
Drafting per-platform versions: automate today. The actual posting-everywhere plumbing: a buy-a-tool question he writes down for later — after Module 2 teaches him how to choose.
Collab-inquiry replies — AI-ASSIST
Frequent but relationship-flavored: AI drafts from his standard answers, Devon reads every one before sending. The doctrine square.
The weekly 'ideas roundup' he sends himself — ELIMINATED
The A/D/E first question caught it: he never reads it back. Deleted outright — 40 minutes a month, zero setup. His favorite verdict of the five.
Module 1 complete — you can see now
Stop and mark it: Module 1's promise is delivered. You can look at your own week and point at exactly what a machine should be doing — and you earned that sentence piece by piece: the bottleneck diagnosis and the Repetition Radar (Lesson 1), the honest line between realistic and hype (Lesson 2), the audit that turned 'busy' into numbers (Lesson 3), and today's decision framework that turned candidates into verdicts. Your Task Audit Worksheet now carries all of it — costed tasks, judgment ratings, and A/D/E verdicts with reasons. That document is the seed of something you'll finish in Module 6: your Automation Blueprint.
What comes next is the part you've been patient for: actually building. Module 2, 'Design It,' takes your top verdict and teaches the craft — how to turn any task into a written recipe, where AI actually goes in a workflow, how to pick tools without drowning in a thousand options, and how to build and test your first real automation. Everything you sorted, audited, and decided this module becomes raw material there. Head back to the academy page and open Module 2 — your 'automate now' square is waiting.
And one habit to carry forever, whatever else you forget: keep the radar on. New repeats will move into your week — they always do. Catch them, cost them, run the sequence: eliminate, delegate, automate. Seeing was always the first skill. Now it's yours.
Key idea: Module 1's promise, delivered: you can look at your own week and point at exactly what a machine should be doing — with numbers and verdicts to prove it. Module 2 teaches you to build it.
What You Can Do Now
You didn't just finish four lessons — you gained working skills. Before you start building, name them, because they're yours now:
You can now:
See repeated work the moment it happens
The Repetition Radar — the bottleneck diagnosis, the Work Log habit, and the reframe: a repetition problem, not a discipline problem.
Tell realistic from hype at a glance
Four families AI handles today · the never-three it should never own · every task marked AI-can, human-must, or mix.
Audit your week into numbers
Evidence over memory, frequency × minutes, queue tasks marked — 'busy all day' became a costed Top 5 with hours attached.
Make the automation decision with a framework
Eliminate → delegate → automate, the worth-it math with its honest costs, and the frequency × judgment matrix.
Keep judgment human from day one
AI drafts, you decide — installed as a reflex before you've built anything, which is exactly when it should be.
Key idea: Come back to this list whenever automation feels abstract again. These aren't topics you covered — they're things you can DO, on your real work, today.
Prompt Lab
Every prompt below is exactly what you saw in the lesson. Copy them, use them, make them yours.
The decision assistant
Your Top 5 with the numbers, this prompt, and ten minutes: verdicts proposed with reasons — for you to confirm, overrule, and own.
Help me run the automation decision on my shortlisted tasks. For each task below I've given: monthly hours, judgment level (low/med/high), whether others wait on it, and my notes. For each one: (1) run the A/D/E sequence — could it be ELIMINATED (would anyone notice if it stopped)? could it be DELEGATED (needs a human, not necessarily me)? only then consider automating; (2) place survivors on the matrix — high/low frequency × high/low judgment — and name the square: automate-now, AI-assist, checklist-it, or keep-manual; (3) sanity-check the worth-it math: estimated setup time vs. monthly savings, including ~10–20% ongoing review attention and a one-line worst-realistic-mistake cost; (4) give a final proposed verdict per task with a one-line reason. Be conservative — recommend 'ignore for now' freely where the math is thin. All final verdicts are mine. MY TASKS: [paste your Top 5 with numbers].
Why it works: The forced A/D/E order catches eliminations that automation-enthusiasm would steamroll past; the matrix naming makes verdicts concrete instead of vibes; and the conservative instruction plus your final say keeps the decision where it belongs — with the person who'll live with the setup and the review.
AI Workflow View
How AI fits into the real-world process — step by step.
See it
Repetition Radar + Work Log (Lesson 1)
Sort it
Four families vs. the never-three (Lesson 2)
Cost it
The audit: frequency × minutes → Top 5 (Lesson 3)
Decide it ★
A/D/E → worth-it math → the matrix square
Build it
Module 2 takes your top verdict from here
Module 1's arc, completed: see it, sort it, cost it, decide it — and the verdicts on your worksheet become Module 2's building list. No tools were needed to get here. That was the point.
Hands-on Lab: Module Capstone: Verdicts on Your Top 5
0/8 steps done
The final activity of Module 1 — your costed shortlist meets the decision framework, and every task leaves with a verdict and a reason.
Your Turn: Do It Yourself
Now it's your turn. Follow these steps and check them off as you go.
You'll know it worked when:
A completed decision worksheet: five verdicts with reasons — typically 2–3 automate/AI-assist, at least one checklist or ignore, and if you're lucky, one glorious elimination. Plus a circled #1 ready for Module 2.
Going further:
Run the A/D/E sequence on ONE task from outside work — household admin, family logistics. If the framework produces a sensible verdict there too, you've internalized it: the decision skill generalizes to every repeated task life sends you.
Download Resources
Knowledge Check
5 questions — click the best answer for each scenario, then check your answers.
Question 1. What's the beginner failure mode this lesson exists to prevent?
Question 2. Why does the A/D/E framework ask 'Eliminate?' FIRST?
Question 3. A task is frequent (daily) but woven with relationship judgment — customer replies that need your read on mood. Which matrix square?
Question 4. The two costs beginners forget in the worth-it math are:
Question 5. Module 1's complete method, in order: (module review)
0/5 answered
Key Takeaways
- The decision fails before the tool does — run the math: savings vs. setup + ongoing attention + mistake cost
- A/D/E in strict order: eliminate first (zero setup, total savings), delegate second, automate what remains
- The matrix makes it mechanical: high-frequency + low-judgment = your first builds; frequent + judgment = AI-assist
- Build-with-AI is the beginner default; buying tools comes later, and written-down ignores don't nag
- 'Not worth it' with a reason is a successful verdict — the goal was the RIGHT automations, never all of them
Lesson Summary
'Is it actually worth the setup?' now has a method instead of a mood: the A/D/E sequence that catches eliminations and delegations first, the worth-it math with its two honestly-counted hidden costs, and the frequency × judgment matrix that names every task's square. Your Top 5 left with verdicts and reasons — and your circled #1 is Module 2's first project. That completes Module 1: you can see repeated work, sort realistic from hype, audit your week into numbers, and decide what deserves automation. The seeing is done. Module 2, 'Design It,' is where you start building — and it's waiting on the academy page.
BrainTrain Pro Tip
Write your verdicts in pen, including the ignores — then don't relitigate them for 90 days. Decision fatigue, not lack of ideas, is what stalls most automation journeys; a verdict with a written reason stays decided, and your energy goes into building the yeses instead of re-debating the maybes.
Module 1: See the Work
0/4 doneYou'll know it worked when:
A completed decision worksheet: five verdicts with reasons — typically 2–3 automate/AI-assist, at least one checklist or ignore, and if you're lucky, one glorious elimination. Plus a circled #1 ready for Module 2.
Going further:
Run the A/D/E sequence on ONE task from outside work — household admin, family logistics. If the framework produces a sensible verdict there too, you've internalized it: the decision skill generalizes to every repeated task life sends you.
