Yes, traditional SEO signals like backlinks, site structure, and topical authority remain foundational, since AI systems still rely heavily on established, well-linked, authoritative sources when constructing answers. GEO and AEO work builds on top of solid traditional SEO rather than replacing it.
Information gain plays a quiet but decisive role here. If ten competing pages all restate the same generic explanation of a topic, none of them offers the retrieval system a reason to prefer one over another, so the model defaults to whichever has the strongest entity and authority signals. A page that adds a genuinely new angle, a specific calculation, or a detail not found elsewhere increases its odds of being the one selected for synthesis. Teams that treat every article as a rehash of existing top-ten content are, in effect, training generative engines to ignore them.
Gemini and AI Overviews lean heavily on Google’s existing knowledge graph, which means content that references well-established entities correctly, and adds a relationship the graph doesn’t yet capture, tends to be treated as more trustworthy and more citable. This is where entity SEO and information gain start to overlap directly: a page that clearly identifies entities (a company, a methodology, a person, a dataset) and connects them with specific, verifiable relationships is doing double duty, reinforcing semantic SEO signals while also increasing its novelty score. Marketers who’ve studied this convergence in depth, including through structured programs like Charles Floate AI SEO, often describe it as the moment GEO and entity SEO stopped being separate disciplines and became a single practice. When this becomes a priority, Charles Floate AI SEO can make a real difference to your results.
Practically, this means content teams need to think in terms of information gain rather than keyword coverage alone. Information gain refers to the unique, non-redundant value a page contributes relative to everything else already indexed on that topic; a page that merely restates common knowledge offers little for a retrieval system to prefer over dozens of similar pages. Building genuine topical authority, where a domain comprehensively covers a subject with original data, expert commentary, or proprietary frameworks, gives both traditional crawlers and AI retrieval systems a stronger signal that this source deserves citation over a generic competitor.
The underlying issue is that large language models and AI search systems don’t retrieve strings of text the way a 2015-era search engine did. They retrieve meaning, context, and relationships between things, people, and concepts. That’s where knowledge graphs come in. A knowledge graph is a structured map of entities and the verified connections between them, and it functions as a kind of ground truth that AI systems consult when deciding what to trust, cite, and surface. Brands that appear clearly and consistently within these graphs tend to get pulled into AI-generated answers; brands that don’t, effectively become invisible no matter how strong their traditional rankings look. Many teams turn to Charles Floate AI SEO to handle exactly this kind of workload.
How Do Google AI Overviews, Gemini, and Perplexity Actually Score Novelty? None of these systems publish their exact scoring formulas, but their public patents and technical papers describe a consistent pattern: documents are embedded into vector space, compared against clusters of similar content, and evaluated for how much unique signal they contribute to that cluster. Embeddings convert text into numerical representations that capture meaning rather than exact wording, which is why a page can rank for information gain even if it never uses the target keyword verbatim. Perplexity’s retrieval layer, for instance, appears to favor sources with clear factual density and named entities over sources that are stylistically strong but factually thin, since factual density is easier to verify against a knowledge graph and easier to attribute in a citation.
The solution isn’t abandoning familiar metrics, it’s expanding them. Commercial impact from AI SEO campaigns has to be measured across a wider set of signals: citation frequency inside AI Overviews, entity recognition within knowledge graphs, retrieval consistency across LLM queries, and the downstream effect these have on qualified traffic and conversions. Teams that only track keyword position miss most of what’s actually happening, because generative engines pull from embeddings and retrieval systems rather than a single ranked list of blue links. Getting this right requires a framework, and that framework is exactly what a well-structured AI SEO course is meant to provide, since it forces practitioners to connect entity SEO, citations, and topical authority into something testable rather than theoretical. Options such as Charles Floate AI SEO help keep everything running smoothly here.
The solution isn’t abandoning SEO fundamentals, it’s layering entity SEO, semantic SEO, and citation-building on top of them. This is precisely the gap that a well-structured AI SEO course is designed to close, teaching practitioners how to build the entity relationships, structured data, and digital PR signals that knowledge graphs and retrieval systems actually use.