Skip to content

03 / Case Study

03Marketing Brain

My marketing strategy, captured as an interlinked skill network — built to take deliverables from 0 to 60 in less time.

Most “AI marketing” tools give you generic best practices. This is the opposite: one operator’s strategic priors, captured.

A semi-hierarchical, interlinked network of Claude skills encoding my unique marketing thinking — so deliverables traverse the cross-domain halo effects (earned ↔ SEO ↔ paid ↔ organic) instead of defaulting to generic best-practice copy.

Problem

Best-practice output is no one’s voice.

Off-the-shelf AI marketing tools ship the consensus answer. Generic frameworks, generic copy, generic strategy — useful as a starting point, useless as a deliverable. Every senior marketer spends the first half of any AI-assisted task wrestling the output away from the average.

The deeper problem: real marketing decisions aren’t isolated. A choice in earned media ripples into SEO, brand, paid, and organic. “Best practice for earned media” without considering those halo effects is wrong by default.

Strategic Approach

Encode the operator. Cross-link the disciplines.

Marketing Brain is a network of Claude skills arranged in a semi-hierarchical tree. At the top, Marketing. Below that, distinct, non-overlapping subcategories — each one a separate skill defining my style and strategy for that slice. Below those, further breakdowns.

The skills cross-link. When you ask for an earned-media strategy, the earned skill pulls from SEO, brand, and advertorials — because that’s how a real marketing decision actually works. The output starts much closer to what I would have written, in my voice, with my strategic priors baked in.

How It Works

Halo-effect-aware skill traversal.

Each skill is a Claude skill file capturing my unique style for a slice of marketing. When invoked, Claude traverses the relevant skills and their cross-links to produce an output that respects the full network — not just the single skill that matches the surface request.

The goal is 0 to 60 in less time: closer to finished on the first shot, skipping the steps where Claude would default to generic advice. Caveat I’ll state plainly: one-shot AI deliverables don’t exist. Human-in-the-loop is still very much a thing. Marketing Brain gets me closer faster — it doesn’t replace judgment.

Demo

See it in motion.

Simulated, zero-token — feels live, costs nothing, can’t be abused.

Loading demo…

Outcomes

What it changed.

  • Deliverables start closer to finished — the network skips the rounds where Claude would default to generic best practice and starts from one operator’s structured thinking, in his voice, with his strategic priors baked in.
  • Cross-links let a request traverse the whole discipline rather than one cell of it: ask for an earned-media strategy and the SEO implications, brand effects and advertorial plays come with it.
  • The caveat, stated plainly: one-shot AI deliverables don’t exist. Human-in-the-loop is still very much a thing — this gets you closer faster, it doesn’t replace judgment.