"AI can't use what you haven't organized." In this solo teaching episode, host Susan Diaz lays out a lightweight, repeatable structure for an internal knowledge base that actually powers your AI - so custom GPTs, Gems, or projects stop guessing and start producing on-brand, accurate work. 

You'll learn the difference between rules (how your AI behaves) and knowledge (what it must know), how to build a four-folder knowledge base, ways to keep it fresh, what not to include for privacy/safety, and a 48-hour challenge to prove it on a real workflow.

What you'll learn

Rules vs knowledge: rules = behaviour, steps, tone, guardrails; knowledge = the factual assets (offers, pricing, voice, proof) your AI must reference. Use both, or you'll get either generic tone or rambling, off-base outputs.

The 4-folder knowledge base: Brand Voice, Product Facts, Policies & Pricing, and Examples - what goes in each, and why this crushes hallucinations.

Freshness rhythm and versioning: set a monthly/bi-monthly review, version by date, and keep a simple changelog so quality doesn't decay.

Privacy and safety notes: what to exclude (confidential contracts, unreleased IP), how to anonymize examples, and who should have edit vs view access.

Live example: how Susan used this exact setup to draft a Northlight landing page that was ~80% right on first pass.

48-hour challenge: create the four folders and drop 1-2 docs into each; test on one real deliverable (do this now)

• Create the four folders.

• Drop 1-2 docs in each (rough is fine).

• Run one real deliverable through your setup; note time saved + edit depth.

• Bring your folder map to Susan's MPC open house for live feedback.

 

Want more?

Agile teams move fast. Grab our 10 AI Deep Research Prompts to see how proven frameworks can unlock clarity in hours, not months. Find the prompt pack here.

Join the Marketing Power Circle (MPC) Connect with Susan Diaz on LinkedIn

Please take a moment to rate and review this podcast: 5⭐ helps more founders find this show