How We Build
Your Second Brain
Four steps. Drop in raw data. AI compiles a wiki. Monthly, it compounds. Here is exactly how it works.
You drop in raw data
Create a folder. Dump everything into it , notes, PDFs, articles, exported docs, old research. No organization needed. The messier, the more impressive the output.
- - Obsidian/Roam/Notion export
- - PDFs and articles
- - Google Docs
- - Notes and journals
- - Meeting notes and client files
Input
raw/
└── obsidian/roam/notion_export
└── pdfs_and_articles
└── google_docs
└── notes_and_journals
└── meeting_notes_and_client_files
AI reads and structures everything
Claude reads every file, identifies concepts, entities, relationships, and themes. It builds a structured schema from scratch , what topics exist, how they relate, which sources back each claim.
- - CLAUDE.md schema generation
- - Concept identification
- - Source citation mapping
- - Cross-reference building
Process
CLAUDE.md
└── claude.md_schema_generation
└── concept_identification
└── source_citation_mapping
└── cross-reference_building
The wiki is compiled
Concept articles are written for every major topic. Source summaries distill each source to its key points. A master index provides navigation. Every article links to related articles and original sources.
- - Concept articles per topic
- - Source summaries
- - Topic index
- - Internal cross-links
Process
wiki/
└── concept_articles_per_topic
└── source_summaries
└── topic_index
└── internal_cross-links
Monthly: it grows and compounds
On the retainer, you drop in new content each month. We compile it into the existing wiki , new articles added, existing ones updated, cross-references extended. Month by month, the system becomes more complete.
- - New content ingested
- - Existing articles updated
- - Health check and audit
- - Q&A reports compiled
Output
outputs/
└── new_content_ingested
└── existing_articles_updated
└── health_check_and_audit
└── q&a_reports_compiled
Month 6 is when it gets good
The value of a second brain isn't at setup. It's in month 6, when you can ask a question and the answer synthesizes three years of reading you'd forgotten you'd done.
The system cross-references topics you didn't know were related. It surfaces a source you read 18 months ago that directly addresses your current problem. That's the compounding loop working.
What Q&A looks like in month 6
Query:
"What are the strongest positioning angles for B2B SaaS buyers in 2024?"
Compiled from 42 relevant sources in your vault:
"Based on your research across the McKinsey 2023 B2B report, your Q3 client debrief notes, and 7 competitive analysis documents... [sourced brief continues]"
Ready to build yours?
Book a setup call and Phillip will walk you through the process, confirm your data scope, and get your second brain built within 48 hours of receiving your files.