Scan time: 2-3 min / Read time: 5-7 min
Hey rebel solopreneurs 🦸♀️🦸♂️
Most creators burn out doing the same tasks over and over.
Smart ones extract the patterns from these manual tasks and let AI handle it forever.
⚡ Why this works
Before you do anything like writing content, research topics, or create outlines for your social media post, you follow a process manually every time.
But you don't document the process. You manually execute every single time. Same questions asked. Same steps taken. Same 4 hours spent.
Here's the thing:
A framework is just a documented process. The questions you ask. The steps you take. The criteria you use to know when you're done.
Most people think frameworks are complex systems only experts create. They're not. Every time you do something repeatedly, you're already following a framework—you just haven't written it down yet.
Your process is like a recipe.
Most creators cook from scratch every single day. They chop vegetables, measure spices, mix ingredients. Every. Single. Time.
It works. But it's exhausting.
Framework creators do something different. They write down the recipe once. Exact measurements. Precise steps. Clear timing.
Then they hand the recipe to someone else.
That "someone else" used to be a human assistant. Now it's AI.
The recipe doesn't change. But you're not in the kitchen anymore. You're the chef who designed the system.
Turns out, getting paid to think beats getting paid to do. Bingo.
Let's see how Nina figured this out:
🔥 This works best with a trained AI sidekick.
Not set up yet? Train in 5 minutes → | Test with sample →
Nina is a newsletter writer. Publishes 3 times a week.
But here's her problem.
After 8 months and 96 newsletters, Nina still spent 4 hours every week on the same research process.
Every topic needed research. Find credible sources. Read 5-7 articles. Extract key insights. Identify unique angles. Organize notes.
Same pattern. Every. Single. Time.
She tried batching research on Mondays. Still 4 hours. She tried templates for note-taking. Still 4 hours. She tried outsourcing to a VA. Explanations took longer than just doing it herself.
16 hours a month on one task.
Nina was tired of being trapped in execution.
Then Nina found something. A principle from creators who'd built million-dollar one-person businesses.
A concept called "Framework Leverage."
It explained exactly why manual execution was killing her scalability. And how to crystallize any process into an AI-executable system in minutes.
Nina decided to follow these steps:
Step 1: Identify her hidden framework Step 2: Train AI to execute it automatically
Nina knew her research process worked.
But she'd never written it down. She just... did it. Open browser. Type topic. Click articles. Read. Take notes. Repeat.
She tried to describe the process to her VA.
Typed: "Research my newsletter topic and find insights."
Too vague. VA returned surface-level stuff.
Typed: "Find 5 credible sources about X topic and extract key points about Y angle."
Better. But still missing something. Which credible sources? What counts as a "key point"? What makes one angle more valuable than another?
Should she list every possible source? Should she define "credible"? Should she explain angle selection?
45 minutes writing instructions. Still incomplete.
The problem? Nina had steps in her head but couldn't see the framework structure underneath.
But if she could extract the questions she asked, the criteria she used, and the outputs she created, she could document the entire system.
Here's what Nina tried:
The framework identifier prompt:
You are a framework extraction specialist.
I do a specific task repeatedly and want to crystallize it into a documented framework.
Your job: Extract the hidden framework structure from my description.
---
ANALYSIS PROCESS:
When I describe my task, identify:
1. **Questions I'm answering** (even if I don't realize I'm asking them)
- What am I trying to determine?
- What decisions am I making?
- What criteria am I using?
2. **Steps I'm taking** (the actual sequence)
- What do I do first, second, third?
- What's the logical order?
- What triggers moving to the next step?
3. **Assets I'm creating** (outputs or tools)
- What documents/files/lists do I produce?
- What templates or structures do I use?
- What's the final deliverable?
4. **Patterns I follow** (rules or principles)
- When do I include vs exclude something?
- How do I know when I'm done?
- What makes a good outcome vs mediocre?
---
RETURN FORMAT:
## Framework Structure: [Give it a descriptive name]
**Hidden Questions:**
- [Question 1]
- [Question 2]
- [etc.]
**Actual Steps:**
1. [Step 1]
2. [Step 2]
3. [etc.]
**Assets Created:**
- [Asset 1]
- [Asset 2]
- [etc.]
**Quality Criteria:**
- [Criterion 1]
- [Criterion 2]
- [etc.]
**One-Sentence Framework:**
[Describe the entire process in one clear sentence]
---
INPUT:
<TASK_DESCRIPTION>
{Describe what you do repeatedly - include what you're trying to achieve, how you currently do it, what you create, and what makes it successful. 2-3 paragraphs.}
Example: "I research newsletter topics every week. I need to find credible sources about my topic, extract unique insights that haven't been covered to death, and identify angles that will make my audience stop scrolling. I usually read 5-7 articles from industry leaders, take notes on surprising statistics or contrarian viewpoints, and organize them by potential newsletter angle. I know I'm done when I have 3-4 unique angles with supporting evidence that feel fresh."
</TASK_DESCRIPTION>
The AI sidekick returned the framework structure.
Framework Structure: The Credible Insight Mining System
Hidden Questions:
Which sources are credible for this specific topic?
What counts as a unique insight vs rehashed advice?
Which angles will resonate with my newsletter audience?
How do I know when I have enough research?
Actual Steps:
Identify 3-5 credible source types for topic
Read sources with filter for "contrarian or surprising"
Extract insights that challenge conventional wisdom
Cluster insights by potential angle
Test angles against audience interests
Select top 3-4 with strongest evidence
Assets Created:
Source list (credible authorities for topic)
Insight database (categorized by angle)
Angle priority ranking (sorted by audience fit + uniqueness)
Quality Criteria:
Credible: Published by recognized expert or backed by data
Unique: Less than 20% of articles say this
Audience fit: Directly addresses my readers' goals
Evidence: Has specific examples or statistics
One-Sentence Framework: "Find 5-7 credible sources, extract insights that contradict mainstream advice, cluster by angle, and select the 3-4 with strongest audience fit and evidence."
Nina read the output.
Wait. That's exactly what she did. Those were the hidden questions. Those were the real steps.
She'd been following a system the entire timeβ€"she just never saw it as a framework.
Completion moment: Nina had her research process documented as a clear, reusable framework.
🤖 Step 2: Nina trains AI to execute it automatically
Nina had her framework structure.
But she didn't know how to turn it into an AI prompt. Should she just paste the steps? Should she explain the questions? Should she give examples?
She tried creating a prompt manually.
Typed: "Research this topic and find unique insights."
Ran it. AI returned generic surface-level stuff.
Tried again: "Find contrarian viewpoints from credible sources about X topic. Extract insights that challenge mainstream advice."
Better. But AI didn't know what "credible" meant for her niche. Returned academic papers when she needed industry practitioner content.
Tried again: "Credible sources means published entrepreneurs with revenue above $1M. Find contrarian insights from 5 sources. Cluster by angle."
Closer. But now the prompt was getting long and confusing. What if she wanted to apply this to different topics? She'd have to rewrite it every time.
30 minutes of prompt iteration. Still not automated.
The problem? Nina had the framework but didn't know the AI training prompt structure that would make it executable.
But if she could generate a complete training prompt from her framework structure, she could paste it once and use it forever.
Here's what Nina did:
The AI training prompt generator:
You are an AI prompt engineering specialist.
I have a documented framework and want to train ChatGPT/Claude to execute it automatically.
Your job: Convert my framework into a complete, executable AI training prompt.
---
CONVERSION PROCESS:
When I provide my framework structure, create a training prompt that includes:
1. **Clear role assignment**
- Tell AI what specialist role to assume
- Define the specific task scope
2. **Framework explanation**
- Explain the core concept
- Define key terms (what makes something "credible", "unique", etc.)
- State the goal outcome
3. **Step-by-step execution**
- Break framework into numbered steps
- Each step = specific action AI takes
- Include decision criteria for each step
4. **Output format specification**
- Exact structure of deliverable
- Sections, lists, or tables to return
- Format that's copy-paste ready
5. **Input section**
- Required fields (what user must provide)
- Optional fields (context that helps but isn't required)
- Examples for each field
---
REQUIREMENTS:
- Make it fully automated (AI does ALL work)
- Include specific decision criteria (not vague like "find good sources")
- Specify exact output format
- Add examples in input placeholders
- Keep under 500 words if possible
- User should just paste topic and go
---
INPUT:
<FRAMEWORK_STRUCTURE>
{Use the framework structure from above}
</FRAMEWORK_STRUCTURE>
The AI sidekick returned the complete training prompt.
(Example output):
You are a credible insight researcher and angle strategist.
Your job: Find contrarian insights from credible sources and cluster them into newsletter angles.
**What "credible" means:**
- Published entrepreneurs/practitioners with proven results (revenue, audience, or track record)
- Data-backed research from recognized institutions
- Industry reports from established companies
- NOT: Generic advice blogs, AI-generated content, or unverified sources
**What "unique" means:**
- Contradicts mainstream advice
- Presents surprising statistics or counterintuitive findings
- Mentioned in less than 20% of articles on this topic
Here's how to execute:
Step 1: Identify 5-7 credible sources for the topic
- Prioritize practitioner content over academic
- Look for sources with demonstrated results
Step 2: Read sources filtering for contrarian or surprising insights
- Flag anything that challenges conventional wisdom
- Note specific examples or statistics
Step 3: Extract insights that meet uniqueness criteria
- Must contradict what "everyone says"
- Must have supporting evidence
Step 4: Cluster insights by potential newsletter angle
- Group similar contrarian viewpoints
- Identify which angles have strongest evidence
Step 5: Rank top 3-4 angles by audience fit + uniqueness score
- Audience fit: Directly addresses target readers' goals
- Uniqueness: Freshness of perspective
Return format:
**CREDIBLE SOURCES (5-7):**
- [Source name + credibility reason]
**CONTRARIAN INSIGHTS:**
- [Insight 1] - Source: [X] - Evidence: [Y]
- [Insight 2] - Source: [X] - Evidence: [Y]
[etc.]
**TOP ANGLES (ranked):**
1. [Angle name] - Uniqueness: X/10, Audience fit: Y/10
- Supporting insights: [List 2-3]
2. [Angle name] - Uniqueness: X/10, Audience fit: Y/10
- Supporting insights: [List 2-3]
[etc.]
---
INPUT:
**Required:**
Newsletter topic: {INSERT e.g., "How to grow on LinkedIn without posting daily"}
Current mainstream advice to avoid: {INSERT e.g., "Post 5x a week, engage for 1 hour daily, follow 50 people"}
**Optional (skip if you've done context setup):**
Target audience: {INSERT e.g., "Busy founders who want LinkedIn presence but don't have time"}
Industry/niche: {INSERT e.g., "B2B SaaS"}
Nina read the training prompt.
Perfect. Every decision criterionβ€"spelled out. Every stepβ€"actionable. Every output sectionβ€"formatted.
She pasted it into ChatGPT/Claude. Added her newsletter topic. Hit enter.
60 seconds later: 7 credible sources, 12 contrarian insights clustered into 4 ranked angles with evidence.
What used to take 4 hours now took 4 minutes.
Completion moment: Nina had an AI sidekick trained to execute her exact research framework automatically.
📊 Nina's results after 4 weeks
Before:
Research time: 4 hours per week
Manual execution: Every single topic
Scalability: Could only research 3 topics weekly
After:
Research time: 15 minutes per week
AI execution: Trained once, runs infinitely
Scalability: Can research 20+ topics weekly
Her process now:
Paste newsletter topic into trained AI (30 seconds)
AI returns credible sources + contrarian insights + ranked angles (60 seconds)
Review output and select best angle (10 minutes)
Use insights in newsletter draft (separate workflow)
Total time: 15 minutes. Not 4 hours.
Her AI sidekick handles the framework execution in 60 seconds. Bingo.
🧠 Your turn
Copy both prompts into your AI sidekick. Run them in the same chat.
Paste your task description into Prompt 1. Your AI sidekick extracts your hidden framework structure.
Then Prompt 2 runs automatically using that framework. It generates your complete AI training prompt.
Generation time: 90 seconds total. Time to have a trained AI employee: 5 minutes.
That's it, my fellow outliers!
Yours 'finding high-potent AI shortcuts so you work less' Vijay peduru 🦸♂️