<div> The emergence of generative AI search engines, including Google AI Overviews, ChatGPT search, and Perplexity, has created a new optimization paradigm: Generative Engine Optimization (GEO). Although practitioners increasingly recognize this shift, the literature still lacks a careful cross-paradigm comparison that separates direct factor evidence from adjacent visibility, overlap, and volatility studies. This paper presents an exploratory comparative synthesis built from a triaged evidence base: a small core quantitative corpus of direct SEO and GEO studies, plus a contextual layer of market, overlap, and citation-behavior sources. The extracted evidence supports three provisional findings. First, GEO-side visibility appears more tightly associated with entity and brand prominence than with classic link-proxy measures. Second, source-credibility interventions such as adding citations and statistics produce substantial gains in the GEO benchmark, suggesting that generative systems reward evidence-bearing and citation-ready content more directly than traditional SEO studies capture. Third, the current overlap literature indicates that GEO still operates on top of an organic search substrate rather than replacing SEO outright. A provisional Divergence Index framework is introduced and applied to the currently matched factor families. Under this framework, authority appears broadly persistent across paradigms, while content-quality divergence remains highly sensitive to construct mapping on the SEO side. The paper therefore argues for a reweighting model rather than a replacement model: familiar factor families persist, new generative-facing signals emerge, and some headline claims about SEO-to-GEO disruption remain stronger than the current evidence base can support. </div>