Generative Engine Optimization (GEO) research has focused exclusively on content-level interventions: adding statistics, citing sources, using authoritative tone, and structuring text for extraction. No peer-reviewed GEO work has proposed infrastructure-level interventions for increasing AI system citation rates. Scattered practitioner recommendations for individual tactics (DOI deposits, ScholarlyArticle schema, repository placement) have emerged independently but remain unconnected, unanalyzed, and untested. This paper proposes the first formal framework for what we term Academic Citation Infrastructure (ACI): the deliberate attachment of scholarly publishing infrastructure to commercial content. ACI unifies four techniques into a coherent strategy: methods-paper formatting, BibTeX/RIS citation file generation, Zenodo DOI mirroring, and ScholarlyArticle schema markup. A review of peer-reviewed GEO research confirms none propose infrastructure-level interventions; a survey of practitioner literature identifies emerging but fragmented adoption. Analysis of 15+ commercial deposits on academic platforms documents the pattern in practice. Quantitative evidence from source framing studies (192,000 assessments), repository discoverability research (+266% search impressions), and LLM search selection analysis (55,936 queries) supports the underlying mechanisms. This paper unifies these scattered tactics into a testable framework, presents the first mechanism analysis explaining why they work, proposes a controlled experimental design for validation, and, as a recursive demonstration, implements all four techniques on itself.
Nuno Andrade (Tue,) studied this question.