# ✅ **MASTER PROJECT CONTEXT — My Complete Lead Gen Automation Blueprint**
I’m building an **n8n-based fully automated personalization engine** for my
Fractional CFO (fCFO) lead-gen agency.
The automation must:
1. **Read my list of leads** from Google Sheet
2. **Decide**:
* Does this lead need *deep research* or just a *normal LinkedIn-level
personalization*?
3. **Detect if the "organizationWebsite" belongs to them or their employer**
* If it’s THEIR firm → Deep website research
* If it’s their employer → Use Tavily/Serper to research THEM specifically
4. **Perform deep research automatically**, using:
* LinkedIn scraped profile (Apify)
* Website scraping & sub-page crawling
* External research tool (Tavily/Serper) when needed
5. **Summarize everything via AI** to produce:
* personal insights
* business context
* industry signals
* credibility markers
6. **Generate personalized icebreakers & subject lines**
7. **Create a Google Doc lead magnet demo**
* Named for the lead (e.g., “Jeff – Deep Personalization Report”)
* Contains an embedded Google Sheet with a 20-lead “demo list”
* The idea: This GDoc becomes the **lead magnet** that shows fCFOs:
> “This is what your startup clients will receive if I build your automation
system.”
8. **Append personalized outputs back into my master CSV/GSheet**
9. **I will manually load these into SmartLead** and fire campaigns
This is the **north star**:
👉 Build a **minimum viable**, fast, reliable automation that produces:
**Deep personalized icebreakers + a polished demonstration GDoc + a SmartLead-ready
data row.**
I do **not** want overengineering.
I want **fast output so I can email TODAY**.
---
# 🧠 **Evolution of the Project (Full Understanding So Far)**
### **1. Started with a huge n8n workflow and asked to understand it**
We broke it down node-by-node.
### **2. Confirmed that:**
* I will NOT scrape LinkedIn again.
* Only company websites will be scraped.
### **3. Built a classification node**
This detects *firm vs employer* and decides how deep research goes.
### **4. Built the “company website path”**
This includes:
1. Scrape homepage
2. Extract links
3. Normalize them
4. Prioritize internal links
5. Fetch top pages
6. Convert to Markdown
7. Summarize each page
8. Aggregate summaries
9. Generate personalization
We are following **Nick’s workflow logic**, but not copying unnecessarily complex
nodes.
### **5. Switched to minimal, fast, clean design**
Because the workflow became long and frustrating,
I insisted on:
* One node at a time
* Only what is needed
* Reusing the existing SplitOutLinks output
* Not overcomplicating link extraction
### **6. Identified the GDoc benchmark requirement**
My automation will create:
* A **personalized GDoc**
* Embedding a **demo Google Sheet**
* Showing what my automation system can do
* This becomes my lead magnet for Fractional CFO clients
### **7. Final Output Must Include:**
For each lead:
* Subject line
* Icebreaker
* Company summary
* One-line snapshot
* Link to GDoc
* Status ("Ready for SmartLead")
---
# 🎯 **Final Workflow Intent (High-Level)**
### **➤ IF the lead owns the firm → Deep Company Website Path**
* Scrape firm URL
* Pull internal subpages
* Summarize all
* Generate icebreaker
* Create GDoc demo
* Append to sheet
### **➤ IF the orgWebsite is employer’s → LinkedIn + Tavily Path**
* Use LinkedIn profile
* Use Tavily/Serper for additional insights
* Generate icebreaker
* Create GDoc
* Append back
---
# 🧩 **Where We Are Right Now in the Build**
We have:
✔ Correct homepage scraper
✔ HTML-to-Markdown conversion
✔ Link normalization
✔ Link splitting
✔ Link prioritization
✔ AI summarizer working
✔ Aggregator (almost done)
✔ LinkUrl attachment fixed
✔ Classification node working
What remains:
🔸 Confirming the aggregator output structure
🔸 Generating a clean multi-page summary
🔸 Generating icebreaker + subject lines
🔸 Creating the GDoc
🔸 Writing the results back to GSheet
We’re very close.
---
# 🚀 **Guiding Principles For ALL Future Steps**
Any assistant working with me MUST:
1. Move **one node at a time**
2. Build **only what moves me toward sending emails TODAY**
3. Not overcomplicate
4. Only add nodes if absolutely required
5. Know the entire above context automatically
6. Maintain continuity — never forget the north star
7. Understand this system is for **fCFO lead-gen personalization**
---
# 🏁 **Final Confirmation**
Yes —
**this summary now includes EVERYTHING: your workflow logic, business intent, lead
magnet strategy, deep research vs shallow path logic, and full technical
direction.**
And yes —
**the entire system is being built with this north star in mind.**
When you’re ready, just say:
**“Continue from next node”**
And we’ll proceed exactly step-by-step toward completion.