Claude Code Just KILLED All Marketing Agencies

Generative Engine Optimization (GEO) Audit Tool with Claude Code

Key Concepts:

• Generative Engine Optimization (GEO): Optimizing websites for AI search engines like ChatGPT, Perplexity, and Gemini, rather than traditional search engines like Google and Bing.
• Claude Code: A platform allowing users to build and run AI-powered tools using “skills” (plugins) and agents.
• Skills: Instructions given to Claude Code to enable it to perform specific tasks or become an expert in a particular area.
• Sub-agents: Specialized AI entities within Claude Code that perform specific analyses as part of a larger task, operating in parallel for efficiency.
• GEO Audit: A comprehensive assessment of a website’s visibility and performance in AI search results.

---

I. Introduction to GEO and the Tool

The video introduces a new approach to SEO called Generative Engine Optimization (GEO), focusing on optimizing for AI-powered search engines. The presenter demonstrates a tool built using Claude Code that automates website audits for GEO, providing a detailed report on a website’s AI visibility. The tool is offered freely for viewers to test and utilize. The presenter highlights the potential of this market, projecting growth to $7 billion in the next few years, driven by the 500%+ year-over-year growth in AI search traffic. Currently, only 23% of marketers are actively considering GEO.

II. Dscript.com GEO Audit Example

A pre-run audit of dscript.com, a video editing tool, is presented as an example. The audit utilized 11 skills and 5 AI agents simultaneously.

• Overall GEO Score: 58/100, indicating significant room for improvement.
• Scoring Categories: The report breaks down performance across AI citability, brand authority, technical content, quality, and schema, each scored individually.
• Platform-Specific Performance: The tool assesses performance on different AI platforms (ChatGPT, Perplexity, Google AI Overviews, Gemini) individually, highlighting potential discrepancies in visibility.
• Critical Findings & Action Plan: The report identifies key issues hindering AI visibility and provides a prioritized action plan for remediation. This report is presented as a deliverable agencies could charge thousands of dollars for, yet was generated in approximately 60 seconds.

III. Live GEO Audit on Typeform.com – Process & Explanation

The presenter then performs a live audit on Typeform.com, a form-building website, demonstrating the tool’s functionality in real-time.

• Accessing Claude Code: The tool is run within a Claude Code instance inside Visual Studio Code (VS Code). Installation is simplified with a single command (link provided in the description).
• Initiating the Audit: The audit is triggered using the /geo-audit command within Claude Code, followed by the target website URL (typeform.com).
• Audit Phases: The audit process is divided into three phases:
  • Phase 1: Discovery & Fetching: Retrieving the homepage and identifying the website’s business type.
  • Phase 2: Parallel Sub-agent Analysis: Running five specialized AI agents concurrently to analyze different aspects of the website.
  • Phase 3: Aggregate Scores & Report Generation: Combining the results from the sub-agents to generate an overall GEO score and report.

IV. Claude Code Skills & Sub-Agent Delegation

The presenter delves into the technical aspects of the tool, explaining the role of “skills” in Claude Code.

• Skills as Plugins: Skills are essentially instructions that give Claude Code expertise in a specific domain. The geo-audit.mmd file contains the detailed instructions for the GEO audit.
• Sub-agent Architecture: Phase 2 utilizes five parallel sub-agents, each focusing on a specific category:
  • AI Citability Analysis: Assessing content for AI readability and codability.
  • Brand Analysis: Evaluating brand recognition and authority.
  • Technical GEO Infrastructure: Analyzing the website’s technical setup for AI crawlers.
  • Content Quality Analysis: Assessing the quality and relevance of the website’s content.
  • Schema Analysis: Evaluating the implementation of schema markup.
• Analogy to Subcontractors: The presenter uses the analogy of a general contractor hiring subcontractors to illustrate how sub-agents work together to complete a complex task efficiently.

V. Typeform.com Audit Results & PDF Report Generation

After completing the audit, the presenter reviews the results for Typeform.com.

• Overall GEO Score: 62/100 (Moderate).
• Score Breakdown: Detailed scores for each category based on the sub-agent analyses.
• Critical Findings:
  • Pricing Page Invisibility: AI crawlers cannot access pricing data due to JavaScript rendering and API calls.
  • Schema Markup Deficiency: Limited schema markup implementation (lack of SoftwareApplication, FAQPage, and Offer schemas).
  • LLM.txt File Non-Compliance: Incorrect formatting of the LLM.txt file.
• Top Strengths: Strong brand recognition (Wikipedia page, LinkedIn followers, G2 ratings) and open crawler access policy.
• PDF Report Generation: A separate skill (geo-report.mmd) is used to generate a professionally formatted PDF report from the audit data. The command /geo-report pdf is used to initiate this process.

VI. Monetization & Community Resources

The presenter discusses potential monetization strategies and provides resources for further learning.

• Value Proposition for Clients: The GEO audit report provides significant value to clients, demonstrating AI visibility issues and offering actionable recommendations.
• Lead Generation Strategy: Presenting the audit report as a free value-add to potential clients, followed by a consultation to discuss improvement strategies.
• Community & Courses: A community is available for learning how to build and sell services based on Claude Code, including courses on starting an AI agency and calendar events for Q&A sessions.
• Installation & Access: The GitHub repository containing the code is freely available (link in description), with a simple one-command installation process. Claude Code can be accessed through VS Code or directly via the claude command in the terminal.

Notable Quote:

“Agencies are charging thousands of dollars for a report like this and they spend a week putting it together. I did this in about 60 seconds with this amazing clot uh code tool that I created.” – Presenter, highlighting the efficiency of the tool.

This tool represents a significant step towards adapting SEO strategies for the evolving landscape of AI-powered search. The ability to quickly and comprehensively assess a website’s AI visibility provides a valuable service for businesses looking to stay ahead in the new era of search.