Mastering Generative Engine Optimization (GEO) with 22 Interactive Tarot Cards

I built GEO Tarot: 22 interactive SVG cards explaining Generative Engine Optimization

UX/UI designer Shinobis IA has released GEO Tarot, a system of 22 interactive SVG cards designed to visualize and explain Generative Engine Optimization (GEO) strategies. The project utilizes a hash-based algorithm and vanilla PHP to demonstrate how websites can optimize for LLM citations rather than traditional link lists.

Why This Matters

While traditional SEO focuses on link ranking, GEO addresses the technical reality of how Large Language Models ingest and synthesize information. By moving beyond isolated pages toward a structured knowledge network using tools like JSON-LD and llms.txt, developers can ensure their technical content is treated as an authoritative source rather than being discarded by LLMs that are increasingly adept at filtering out generic AI-generated filler content.

Key Insights

• The Hierophant (V) introduces llms.txt, a root-level file that functions as the robots.txt for generative AI models (2026).
• The Emperor (IV) utilizes JSON-LD Knowledge Graphs to connect content via relatedLink and about properties, turning blogs into machine-readable networks.
• The High Priestess (II) concept of Sentiment Mapping highlights how LLMs distinguish between authoritative assertion and speculation to determine citation weight.
• The Devil (XV) represents generic AI content, which LLMs are now programmatically identifying and deprioritizing in generative answers.
• The Star (XVII) represents the metric of a direct ChatGPT citation, validating the implementation of citable content structures over traditional traffic metrics.

Practical Applications

• Use Case: Deploying Cards IV (Knowledge Graph) and V (llms.txt) to immediately improve a site’s visibility to AI crawlers; Pitfall: Failing to provide a citable content structure (Card III), which leaves LLMs unable to extract clean references.
• Use Case: Applying Sentiment Mapping (Card II) to technical documentation to increase authority scores; Pitfall: Publishing unedited AI-generated articles (Card XV), which decreases the overall weight given to the domain by generative engines.

References:

• https://dev.to/shinobis_ia/i-built-geo-tarot-22-interactive-svg-cards-explaining-generative-engine-optimization-1jbe