AI Advantage Audit

Install in one command

mkdir -p ~/.claude/skills/ai-advantage-audit && curl -fsSL https://eliteaiadvantage.com/skills/ai-advantage-audit/SKILL.md -o ~/.claude/skills/ai-advantage-audit/SKILL.md

Walk through the 5 lenses of "The AI Advantage Audit", Offer, Audience, System, Brand, State, and produce a personalized report with the top 3 AI moves to make next, ranked by ROI window. Same framework as the white paper at /white-papers/ai-advantage-audit and the form at /audit, but interactive and conversational.

The cross-cutting thesis: AI is a force multiplier on whichever business layer is already strong, and a magnifier of whichever layer is broken. So the audit is not "where can we add AI?", it is "which layer do we stabilize first so the AI we add later actually pays?"

Built for SMB owners, mid-market marketing/ops directors, and consultants who need to make a real call about where to deploy AI capital. Not generic advice, every recommendation cites your actual answers.

• →Forces a real diagnostic instead of vibe-driven AI bets, every recommendation cites a specific answer you gave.
• →Surfaces where AI would multiply vs. magnify, same move can compound or destroy depending on the lens.
• →State-aware: gives a survival-mode business different advice than a scaling business, even with similar answers.
• →Stops you from building the wrong AI thing first, usually the move you assumed was right.
• →Bridges to action with 3 prioritized AI moves with ROI windows, not a generic "consider these areas" list.

• •You are deciding where to invest AI build hours or budget next.
• •You inherited an AI strategy and need to pressure-test it.
• •You are auditing a client business before pitching an engagement.
• •You are about to spend on an AI tool and want a sanity check.
• •You are stuck on an AI plateau and the "obvious" next move feels off.

Provides Claude with the explicit 5-lens framework (Offer / Audience / System / Brand / State), the multiplier-vs-magnifier scoring rubric, and state-specific defaults (survival / plateau / growth / scaling). Without this skill, Claude tends to give generic AI advice ("consider automating customer support"). With this skill, Claude collects 15 specific answers, classifies each lens STRONG/MIXED/BROKEN, and outputs a prioritized report that cites the actual answers. The skill enforces voice rules (no jargon, short paragraphs, honest verdicts) and structural rules (top 3 moves only, each with why-now / what-to-build / ROI window). The model is calibrated to push back on vague answers ("we are flexible") rather than accept them.