Generative Engine Optimization (GEO): How to Optimize Websites for LLMs, AI Agents, and the Next Generation of Search Engines
Traditional SEO was designed to optimize websites for search engines that rank links and display content for humans.
With the emergence of large language models, AI assistants, and agent-based systems, this paradigm is undergoing a fundamental shift.
Modern AI systems no longer just follow links; instead, they extract structured information, interpret content semantically, and generate answers from it—often without users ever visiting the original website.
This presentation introduces the concept of Generative Engine Optimization (GEO)—a technical approach to designing websites so that they can be correctly read, understood, and utilized by AI systems.
Topics covered include:
- How LLM-based systems crawl and process web content
- Why semantic structure is becoming more important than keywords
- Use of Schema.org, JSON-LD, and structured data
- Architectural patterns for AI-readable websites
- Differences between classic SEO, API-First, and GEO
- Preparing existing applications for AI-powered search
- Practical examples and implementation approaches
This presentation has a technical focus and is aimed at developers, software architects, and engineers who want to prepare their applications for the next generation of search, AI interfaces, and agent systems.
This is not a marketing talk, but a technical look at how AI is changing the web.
Gianluca Roth
