How to serve content to agents (a field guide)

The discussion around optimizing content for AI citations explores the emergence of Agent Engine Optimization (AEO), which involves making content accessible and attractive for AI models to cite and recommend. This process is divided into two key areas: positioning, which focuses on how content is presented to AI models during training and retrieval, and content consumption, which involves serving content effectively when AI agents access it. Despite ongoing debates and research by companies like Profound, there is no definitive strategy for increasing AI citations, as the landscape is rapidly evolving, with AI models improving their content processing capabilities. Strategies like using markdown, implementing llms.txt files, and serving content via APIs or markdown routes are examined, each with its own benefits and limitations. The emphasis is on maintaining clean content structures and standards-based content negotiation, which are durable investments that will remain relevant regardless of changes in AI technology. Ultimately, the goal is to serve quality content in appropriate formats to meet both human and AI needs, acknowledging that while the impact of format on AI citations remains uncertain, structured content and robust infrastructure provide a solid foundation for future developments.