The Bayesian Content Optimizer runs a four-stage iterative loop—comprising a Seeder, an Acquisition policy, an Ideator, and an LLM Ranker—to systematically identify, test, and apply content modifications that improve a target URL's ranking position in model-generated search responses. Operating across both snippet and full-page modes, it samples hypotheses modeled as Beta distributions, uses an LLM Ideator to execute precise text edits against an active Rank Factor Class taxonomy, and scores the resulting variants via an LLM judge to deterministically update posterior beliefs until reaching an optimal rank.