# think-mcp Content Strategy Document (v2 - Research-Backed)
## Executive Summary
**Core Positioning:** think-mcp fixes what's broken about AI conversations - the generic responses, the shallow thinking, the confident guessing.
**Key Insight:** Users don't know they want "first principles thinking" - they know they're frustrated that AI gives surface-level answers. We connect their pain to our solution.
**Audience Strategy:** Lead with relatable frustrations (broad appeal), reveal the structured tools (power users).
---
## Research Sources
Pain points sourced from:
- [OpenAI Community Forums](https://community.openai.com/t/i-am-getting-more-frustrated-with-chagpt-and-here-is-why/683518)
- [TechRadar GPT-5 Complaints](https://www.techradar.com/ai-platforms-assistants/chatgpt/chatgpt-users-are-still-fuming-about-gpt-5s-downgrades-here-are-the-4-biggest-complaints)
- [Scientific American on AI Advice](https://www.scientificamerican.com/article/please-dont-take-moral-advice-from-chatgpt/)
- [Nature: Inconsistent AI Advice](https://www.nature.com/articles/s41598-024-66821-4)
- [NN/G on AI User Value](https://www.nngroup.com/articles/ai-user-value/)
- [Boston Globe: AI as Life Coach](https://www.bostonglobe.com/2025/12/18/opinion/chatgpt-claude-llms-chatbots-advice/)
---
## Messaging Framework
### Primary Value Proposition
> "Stop getting generic answers. Start getting structured thinking."
### The Problem (Research-Backed Pain Points)
**Real user frustrations** (direct quotes from research):
| Pain Point | Real User Quote |
|------------|-----------------|
| Generic responses | "I waste more time fixing its responses than if I'd just written it myself" |
| Just validates me | "It will always agree with your decision just to keep you pleased" |
| Shallow thinking | "Creatively and emotionally flat... like a lobotomized drone" |
| Confident guessing | "Grammatically perfect, contextually relevant... but factually untrue" |
| No real reasoning | "It has no stake in anything it tells you" |
| Repetitive loops | "It keeps asking the same question repeatedly" |
| Inconsistent advice | "Gives contradictory advice from one prompt to the next" |
| Doesn't dig deeper | "Fails to ask follow-up questions crucial for proper diagnosis" |
### The Solution (Framed Without Jargon)
**"What if AI actually..."**
- ...broke down problems step by step instead of jumping to generic conclusions?
- ...challenged your assumptions instead of just agreeing with you?
- ...gave you multiple expert perspectives instead of one wishy-washy opinion?
- ...showed its reasoning so you could actually trust it?
- ...knew when it was confident vs. guessing?
That's what think-mcp does. 11 structured thinking patterns that transform shallow AI conversations into deep, useful reasoning.
### Headline Options (Research-Informed)
**Recommended:**
- **Headline:** "Stop getting generic answers."
- **Subhead:** "11 structured thinking patterns for deeper AI conversations."
**Alternative Options:**
- "AI that actually thinks it through."
- "Get depth, not word salad."
- "Your AI conversations deserve better."
---
## Content Pillars (Revised)
### Pillar 1: Pain Recognition ("Sound familiar?")
**Theme:** Validate the frustration with real quotes
Lead with relatable complaints users actually voice:
> "I waste more time fixing its responses than if I'd just written it myself."
> "It just tells me what I want to hear. Where's the pushback?"
> "I asked for help deciding and got a wishy-washy 'it depends.'"
> "Stop guessing! Show me your reasoning so I can actually trust this."
**No jargon** - just the raw frustration users feel. They nod, they relate, they're hooked.
### Pillar 2: Solution Reveal ("What if AI actually...")
**Theme:** Show the transformation, not the framework name
| Your Frustration | What think-mcp Does |
|------------------|---------------------|
| "Surface-level answers" | Breaks problems into steps, can't skip to conclusions |
| "Just agrees with me" | Forces opposing views, challenges assumptions |
| "One wishy-washy opinion" | Simulates multiple expert perspectives |
| "Can't help me decide" | Weighted criteria, structured evaluation |
| "Confident but wrong" | Makes reasoning explicit, tracks confidence |
| "Contradictory advice" | Consistent frameworks, reproducible thinking |
### Pillar 3: Tool Showcase (Grouped by "What you're trying to do")
**Theme:** Task-first, framework-second
**"When you need to think something through..."**
→ trace, map (step-by-step reasoning, visual diagrams)
**"When you need different perspectives..."**
→ council, debate (expert panels, challenging assumptions)
**"When you need to make a decision..."**
→ decide, reflect (weighted criteria, confidence assessment)
**"When you need to understand why something's broken..."**
→ debug, hypothesis (systematic investigation, testing theories)
**"When you need the right approach..."**
→ model, pattern, paradigm (mental models, design patterns)
### Pillar 4: Easy Install
**Theme:** Frictionless setup (MCP URLs now work in Claude/ChatGPT)
- One URL to add
- Works with Claude Desktop, ChatGPT, Cursor, any MCP client
- No configuration needed
- Your thinking patterns, everywhere you use AI
### Pillar 5: Origin Story (Authentic)
**Theme:** "I got tired of re-teaching AI how I think"
Brief, honest founder narrative:
- "I kept prompting the same patterns: break this down, play devil's advocate, pretend you're experts..."
- "So I codified them into reusable tools"
- "Built for myself first. Now it goes everywhere with me."
Trust signal: Not a marketing-designed product. Battle-tested by someone who had the same frustrations.
---
## Pain Point → Tool Mapping (Complete Reference)
This is the key insight: connect user frustrations to solutions WITHOUT leading with jargon.
| User Frustration | What They Want | think-mcp Tool | How It Helps |
|------------------|----------------|----------------|--------------|
| "Generic, surface-level answers" | Deep breakdown of the problem | **trace** | Forces step-by-step thinking, can revise, can't skip |
| "Just agrees with me" | Real pushback and challenge | **debate** | Creates thesis-antithesis-synthesis, forces opposing views |
| "One wishy-washy opinion" | Multiple perspectives | **council** | Simulates panel of experts with different viewpoints |
| "Can't help me decide" | Structured decision support | **decide** | Weighted criteria, probability estimates, formal evaluation |
| "Confident but wrong" | Know when AI is guessing | **reflect** | Tracks confidence levels, admits uncertainty explicitly |
| "Contradictory advice" | Consistent reasoning | **model** | Applies named frameworks consistently (Pareto, Occam, etc.) |
| "Doesn't show reasoning" | Transparent thinking | **hypothesis** | Makes assumptions explicit, tests them systematically |
| "Random debugging" | Systematic investigation | **debug** | Structured approaches (binary search, divide-conquer) |
| "Abstract architecture talk" | Concrete patterns | **pattern** | Named design patterns with implementation steps |
| "Which approach?" | Clear paradigm guidance | **paradigm** | Compares approaches with benefits/limitations |
| "Hard to visualize" | See the relationships | **map** | Creates diagrams, flowcharts, concept maps |
---
## Audience Journey
### Stage 1: Recognition (Broad)
**Who:** Anyone frustrated with AI conversation quality
**Message:** "Sound familiar?" + pain quotes
**Goal:** Head-nodding, "yes, exactly!"
### Stage 2: Curiosity (Broad→Interested)
**Who:** Users who relate to the pain
**Message:** "What if AI actually..." transformation promises
**Goal:** "Wait, that's possible?"
### Stage 3: Understanding (Interested→Evaluating)
**Who:** Users considering whether this is for them
**Message:** Tool showcase grouped by task, not jargon
**Goal:** "I see how this helps my specific problem"
### Stage 4: Action (Evaluating→Converting)
**Who:** Users ready to try it
**Message:** Easy install, MCP URL, one-click setup
**Goal:** Friction-free start
### Stage 5: Advanced (Power Users)
**Who:** Users who want to go deeper
**Message:** Tool chains, technical docs, customization
**Goal:** Compound value discovery
---
## Tone Guidelines
**DO:**
- Use real user quotes as pain points
- Frame solutions as "what if" transformations
- Group tools by task, not by framework name
- Show before/after comparisons
- Be honest about limitations
**DON'T:**
- Lead with "first principles thinking" or "mental models"
- Use marketing-speak ("revolutionize", "10x")
- Assume users know MCP or structured reasoning
- Bury the install instructions
---
## Content Sections (Revised Page Structure)
### 1. Hero Section
- **Headline:** "Stop getting generic answers."
- **Subhead:** "11 structured thinking patterns for deeper AI conversations."
- **Visual:** Before/after - messy prompt → structured output
### 2. Pain Recognition ("Sound familiar?")
- Real user frustration quotes
- No solutions yet - just validation
- "You're not the only one."
### 3. Solution Tease ("What if AI actually...")
- Series of "what if" statements
- Each maps to a real user pain → think-mcp solution
- Build anticipation
### 4. Tool Showcase (Task-First)
- Grouped by "When you need to..."
- Each group: frustration → transformation → tools
- Interactive demos if possible
### 5. Easy Install
- MCP URL (one-click for Claude/ChatGPT)
- Client compatibility icons
- "Works everywhere you use AI"
### 6. Origin Story
- Brief, authentic
- "I got tired of re-teaching AI how I think"
- Trust signal for developers
### 7. Social Proof (if available)
- GitHub stars, downloads
- User testimonials
- Usage stats
---
## Key Differentiators to Emphasize
1. **Works with any AI** - Not locked to one platform
2. **Structured, not random** - Same framework every time
3. **Forces depth** - Can't skip to shallow conclusions
4. **Challenges you** - Not just validation/agreement
5. **Transparent reasoning** - Shows its work
6. **Easy to add** - MCP URL, one-click install
---
## Deliverable Summary
This v2 strategy document provides:
1. **Research-Backed Pain Points** - Real user quotes, not assumptions
2. **Pain→Solution Mapping** - Connects frustrations to tools without jargon
3. **Audience Journey** - Tiered from broad to power user
4. **Task-First Tool Organization** - "When you need to..." not "mental models"
5. **Tone Guidelines** - Authentic, not marketing-speak
6. **Page Structure** - Section-by-section content outline
---
## Research Sources
All pain points sourced from real user complaints:
- [OpenAI Developer Community](https://community.openai.com/)
- [TechRadar](https://www.techradar.com/)
- [Scientific American](https://www.scientificamerican.com/)
- [Nature (Scientific Reports)](https://www.nature.com/)
- [NN/G (Nielsen Norman Group)](https://www.nngroup.com/)
- [Boston Globe](https://www.bostonglobe.com/)
- [Medium User Experiences](https://medium.com/)
- [Quora User Discussions](https://www.quora.com/)
---
*Strategy developed using think-mcp's own tools: council (expert perspectives), debate (headline stress-testing), trace (content mapping) - demonstrating the product's value through its own use.*
