Google’s LLM patent suggests a new goal for SEO: Teaching AI who you are

Google filed a patent in 2023 detailing a system where Large Language Models (LLMs) can build a comprehensive understanding of entities like businesses, brands, and products by analyzing websites and public data. This "Data extraction using LLMs" patent outlines a process for extracting information, identifying relationships between data points, and synthesizing this into a "deep, holistic characterization" of an entity. The technology aims to move beyond simply indexing documents to understanding the underlying entity represented by the content. This shift suggests that Search Engine Optimization (SEO) may evolve to focus on helping search engines like Google comprehend the entity behind the content, rather than solely optimizing the content itself. As search products become more conversational and recommendation-driven, a deep understanding of entities will be crucial for AI systems to effectively recommend businesses, compare products, explain brands, or suggest service providers. The patent defines an entity broadly, encompassing people, companies, businesses, places, objects, and concepts, indicating a move towards a more nuanced interpretation of information and relationships.