Before we install anything, understand what you're actually building. Not just a coaching program. A knowledge operating system that makes you 10x faster — and gets better the more you use it.
You've probably used ChatGPT or Claude. You typed a question. You got an answer. The answer was OK. You closed the tab. Next session, you started over — and the AI remembered nothing.
That's not a flaw in the AI. It's a missing layer. The AI has broad general knowledge but zero specific knowledge about you, your clients, your voice, your business. Unless you give it that context explicitly, every output is generic by default — because it has nothing to draw from but the average of everything it was trained on.
Generic LinkedIn template. Sounds like every other consultant. You re-explain your business, your buyer, your tone in every prompt — and the AI still produces something that could come from anyone.
Specific to your buyer's situation. Uses their language from your pain maps. Sounds like you wrote it. Produced in minutes. Ready to send with minor edits.
The difference isn't the model. It's the context the model has to work from.
A knowledge library is a structured set of files that describe your business — your positioning, your buyer, your voice, your frameworks, your decisions. When you sit down to do real work with AI, you load the relevant files. The AI reads them. Now it knows your business — and produces output that's grounded in your reality, not generic averages.
This is what you're building over the next 90 days. Not just positioning docs. Not just an offer document. A structured knowledge system that any AI can load to act like a partner who's been embedded in your business for months.
In the next module (0.2), you'll install the Strategist Coach — your primary AI partner — and load the skills that will produce your knowledge library file by file. The first skill you'll invoke (in Phase 1) is the Compound Diagnostic, which produces your first canonical file. By Day 90, you'll sit down with your AI, load five files, and produce work in minutes that used to take hours. By Year 2, the system runs your business in 2 hours a week.
As your knowledge library grows, you'll naturally have different types of files doing different jobs. Knowing the difference helps you maintain the library without it turning into a mess.
| Type | Purpose | Examples |
|---|---|---|
| Canonical | The source of truth on a topic. Permanent. Updated when reality changes. | positioning.md, offer.md, compound-consultant-diagnostic.md |
| Working | Active context for an ongoing project or decision. May get promoted to canonical or archived. | q2-planning-working-notes.md, sales-page-draft.md |
| Reference | External info you want to remember. Research, frameworks, articles, examples. | competitor-research.md, pricing-models-reference.md |
| Archive | Old versions, completed projects, decisions that have run their course. | 2025-q4-priorities-archive.md |
Your diagnostic (Phase 1.1) will be your first canonical file. Phase 1 produces more canonical files (positioning, ICP, core problem, pain maps, desire maps, narrative spine). Phase 2 adds more (offer doc, sales page, voice guide). The library grows as your business does.
You don't need to organize this rigidly now. Your Strategist Coach (which you install in Phase 2) will help maintain the library — creating files when domains emerge, updating files when decisions or signal demand it, promoting working notes to canonical when patterns stabilize, archiving stale files.
Every artifact you create over 90 days is a knowledge asset. By itself, each file is useful. The compounding happens because each file makes the next one sharper — and over time the system becomes worth more than any single output.
Your diagnostic. Useful as a baseline, but limited on its own. AI can read it and understand your starting point — that's it.
Diagnostic + six Phase 1 positioning files. Now AI can write outreach in your voice, draft proposals that speak to your buyer's actual situation, prep discovery calls in minutes.
Positioning, offer, sales page, voice guide, pipeline systems, content strategy, decision log. The AI knows your business deeply. You produce in minutes what used to take hours — and the system runs in 2 hours a week of maintenance after that.
This is what "compounding" actually means in this context. Most consultants iterate piecemeal — they tweak a LinkedIn post, rewrite an email, redo a proposal. Nothing accumulates because there's no source to update. The work is repeated, not built on.
You're iterating at the source. Every market conversation feeds back into the canonical files. Every revision improves every future output. The library doesn't depreciate like software does — it appreciates because it captures more of your real business over time.
At the start, your knowledge library is just a folder of files. By the time you reach Phase 2, it becomes the foundation for something bigger — an AI operating system that runs strategic thinking, decision-making, and production work across your entire business.
You'll install your Strategist Coach in the very next module (0.2) — your primary AI partner that reads every canonical file at the start of every session. It diagnoses where you are, recommends what to work on, and runs specialized skills (Compound Diagnostic, Positioning Architect, Offer Architect, more to come) for structured production work. When work belongs in a separate project — content production, IP development, automation — the Strategist scopes a brief you take with you.
You don't need to understand all of this now. The point is to know what you're moving toward: one strategic AI partner + multiple skills + canonical files + project briefs for separate work. The Compound O.S. is the architecture above all your production tools. It's what makes the system durable and the value compound.
You don't need to build a complex folder structure today. You need two things.
/Compound Consultant OS. That's the whole setup. No subfolders required yet. Your canonical files will live here as you produce them, starting in Phase 1.