Free playbook · No upsell

How I turned 30 hours of my own data into a personal AI Chief of Staff

The full playbook. Every script. Every prompt. Every file structure for Claude, ChatGPT, and Obsidian. Drop your email and I'll send it.

You will also get occasional notes from Cameron on building MCG. Unsubscribe anytime.

Six phases. ~1,800 words. Every script you need.

No fluff. No "AI will change everything." Just the exact build I did over 30 hours, in the order you should do it.

  1. PHASE 1

    The 12 data sources to export. ChatGPT, Apple Notes, Gmail, iMessage, LinkedIn, Apple Health, Whoop, calendar, transactions, medical records, handwritten journals. Exact export paths for each.

  2. PHASE 2

    The persona doc prompt. The single highest-leverage prompt in the entire system. This is what makes the AI sound like it has known you for two years instead of two minutes.

  3. PHASE 3

    Image OCR for handwritten pages using Claude Vision. Apple's OCR mangles cursive. Claude reads it. Under $2 to OCR 89 pages.

  4. PHASE 4

    The LinkedIn rolodex builder. Turns your Connections.csv into a "who should I reach out to about X" engine. Grouped by company, role, industry, and recency.

  5. PHASE 5

    The install. Exact folder structures for Claude Projects, ChatGPT Projects, and Obsidian. The global instruction files for each surface. The retrieval map that tells the AI which file to look at for which question.

  6. PHASE 6

    How to maintain it. What to refresh quarterly. When to update the persona doc. When to fold a recurring correction into your global instructions.

Three kinds of people get the most out of this.

  • 01

    Fractional CFOs and finance advisors who want to productize a client operating system, not reinvent it for every engagement.

  • 02

    Founders running two or more projects who want an AI that knows the full context, not just the current chat.

  • 03

    Anyone who has uploaded a ChatGPT history to a new project, gotten nothing useful back, and assumed the model was the problem. The model is not the problem. Structured context is.

Cameron Moore is the founder of Moore Capital Group, a fractional CFO firm serving owner-led businesses in construction, home services, property management, and local services.

He built this second-brain system over 30 hours after being told by every productivity guru that AI was going to change everything, and finding that nothing actually worked until he changed how he fed it context.

More about Cameron →

Cameron Moore

Want this built for your firm?

The same system, applied to your client operating model.

If you run a fractional CFO practice, a consulting firm, or an advisory business and you want to productize this as part of your client onboarding, that's what I do at MCG.