THE STEPS
Step 1: Identify the “canteen task” you’re protecting (5 minutes)
In prison economies, money barely matters. What matters is access—to food, favors, protection. People hoard what gives leverage. That hoarding creates inefficiency, but also safety.
You’re doing the same thing with your work.
Action:
Open your notes doc. Write the answer to this exact question:
What task do I refuse to delegate because “only I can do it right”?
Be specific. Not “strategy.” Write:
- “Writing first drafts of client emails”
- “Turning messy notes into summaries”
- “Deciding what to work on next”
What to expect:
Mild defensiveness. A story about quality. Ignore it.
Common mistake to avoid:
Choosing a task you wish you did more often. Pick one you actually do.
I picked email drafts. I thought they were my edge. They were my candles.
Step 2: Time-box your current method (7 minutes)
Before electricity, people optimized candles endlessly. Better wax. Better wicks. Same ceiling.
You need a baseline.
Action:
Do the task once, your normal way. Time it. Don’t rush.
Write down:
- Start time
- End time
- How drained you feel (1–10)
For me:
- 18 minutes
- Drain: 6/10
- Quality: “fine”
What to expect:
A little boredom. That’s data.
Common mistake to avoid:
Multitasking. This ruins the comparison.
Step 3: Describe the task like you’re trading it (6 minutes)
Back to prison economics.
Trust is currency. Reputation enforces it peer-to-peer. If you want someone else to do a job, you don’t say “do it well.” You specify rules, boundaries, consequences.
AI works the same way. Vibes fail. Contracts work.
Action:
Copy-paste this prompt into your AI tool:
I want you to take over a task I normally do myself.
First, ask me questions to clarify:
- The goal of the task
- The audience
- What a “good result” looks like
- What mistakes would be unacceptable
Do NOT perform the task yet. Only ask questions.
Answer the questions honestly. Short sentences.
What to expect:
The AI will ask 5–8 questions you’ve never explicitly answered before.
Good.
Common mistake to avoid:
Over-explaining. Treat this like onboarding a competent coworker, not a child.
Step 4: Install the “electricity switch” prompt (8 minutes)
Here’s where most guides fail. They show you a clever prompt. That’s a candle.
You’re going to create a standing instruction—a reusable mental grid.
Action:
After answering the questions, paste this:
From now on, whenever I give you input related to this task, follow these rules:
1. Assume you are the primary owner of the task.
2. Produce a complete first version without asking permission.
3. Explain your reasoning briefly at the end.
4. Flag any uncertainty instead of guessing.
Acknowledge these rules, then wait.
Wait for the acknowledgment.
What to expect:
A short confirmation. Nothing flashy.
This is infrastructure.
Common mistake to avoid:
Tweaking the wording endlessly. Clarity beats cleverness.
(If you don’t want to spend weeks refining prompts like this, there are battle-tested prompt packs at wowhow.cloud/products that shortcut the trial-and-error. I resisted this at first. Pride is expensive.)
Step 5: Run the task through the system (7 minutes)
Now flip the switch.
Action:
Paste the raw input you’d normally work from.
For email drafts, I pasted bullet points. For you, it might be:
- Meeting notes
- A client request
- A rough idea
Do not add instructions unless absolutely necessary.
What to expect:
A complete output. Not perfect. Surprisingly usable.
Mine was 80% there. Which bothered me. I’ll come back to that.
Common mistake to avoid:
Jumping in mid-generation to “fix” it. Let it finish.
Step 6: Enforce reputation, not micromanagement (6 minutes)
In prison systems, reputation sticks. You don’t relitigate every trade. You adjust once, then remember.
Do the same.
Action:
Give feedback in this format only:
Feedback:
- Keep doing:
- Do less of:
- Do more of:
No rewriting. No emotional language.
What to expect:
The second version improves sharply. Mine jumped from 80% to ~92%.
Common mistake to avoid:
Editing the output yourself before feedback. That teaches nothing.
Step 7: Compare the economics (6 minutes)
Now the uncomfortable math.
Action:
Time how long the AI-assisted version took including feedback.
Write down:
- Total time
- Drain level
- Quality vs. your original (honest)
My numbers:
- 11 minutes
- Drain: 2/10
- Quality: better than mine on structure, slightly worse on tone
Here’s the contradiction:
AI was everything. Except tone. So I kept tone.
That’s electricity. You don’t generate it. You route it.
Common mistake to avoid:
Demanding 100% replacement. That’s candle thinking.
Step 8: Lock the system (5 minutes)
Most people stop here. Then they forget. Candles again.
Action:
Save the prompt + rules in one place:
- A pinned chat
- A doc titled “AI: [Task Name]”
Add one sentence at the top:
This replaces my old method.
That sentence matters more than it should.
What to expect:
A weird sense of relief. And loss.
Common mistake to avoid:
Treating this as an experiment instead of a migration.
Because they treat AI like a favor, not a currency system.
In prison economies, hoarding trust kills scale. Sharing it—with rules—creates networks. Most people never make that mental shift with AI tools for beginners. They ask for help. They don’t grant authority.
I argued against this idea for days. Surely prompts were the issue. Or model quality. Or my inputs.
What survived the attack: ownership.
The moment I made the AI the primary owner of the task, everything changed. Not magically. Structurally.
THE RESULT
Here’s what my finished setup looks like:
- One saved prompt that handles email drafts end-to-end
- A consistent feedback loop that improves outputs without rewrites
- A reclaimed 7 minutes per email × ~12 emails/week = 84 minutes
- Less cognitive drain (this was the real win)
I don’t “use” AI for this anymore.
I rely on it.
LEVEL UP
Once this feels boring (it will), do one of these:
-
Chain tasks
Feed the output of one AI-owned task into another.
-
Add constraints
Word limits. Tone rules. Risk flags.
-
Transfer reputation
Tell the AI: “Apply the same standards you use for Task A to Task B.”
-
Audit monthly
10 minutes. What drifted? Fix once.
Electricity didn’t replace candles overnight. It replaced the need to think about light.
That’s the real ai productivity transformation. And yes—this is how to use AI effectively without turning into a prompt goblin.
Candles feel safe. Electricity scales.
Choose.
Want to skip months of trial and error? We've distilled thousands of hours of prompt engineering into ready-to-use prompt packs that deliver results on day one. Our packs at wowhow.cloud include battle-tested prompts for marketing, coding, business, writing, and more — each one refined until it consistently produces professional-grade output.
Blog reader exclusive: Use code BLOGREADER20 for 20% off your entire cart. No minimum, no catch.
Browse Prompt Packs →
Share this with someone who needs to read it.
#AIforBeginners #AIProductivity #PracticalAI #PromptEngineering #WorkflowDesign #DigitalLeverage
Comments · 0
Beta: comments are stored locally on your device and not visible to other readers.
No comments yet. Be the first to share your thoughts.