Casper Media has released findings from a large-scale benchmark study measuring how brands are cited across three major generative AI platforms -- ChatGPT Search, Perplexity, and Google AI Overviews -- marking one of the more methodologically detailed examinations of AI search visibility conducted to date. The study, built on more than 5,000 structured query prompts, analyzed citation behavior across 500 enterprise brands and 25,000 local business profiles, producing what the company is calling the Brand Citation Benchmark.
Study Scope and Methodology
The Brand Citation Benchmark was designed to evaluate how consistently generative AI platforms surface, reference, and attribute brands in response to informational and transactional queries. Casper Media structured the prompt set to mirror realistic search behavior across industries, geography, and query intent. The 5,000 prompts were distributed across enterprise and local business categories, enabling a side-by-side comparison of citation frequency and accuracy on ChatGPT Search, Perplexity, and Google AI Overviews.
The local business segment of the study placed particular emphasis on NAP consistency -- the alignment of a business's Name, Address, and Phone number across authoritative data sources. Discrepancies in NAP data correlated with measurably lower citation rates across all three platforms, a finding that carries direct implications for businesses operating in competitive local markets. Among the 25,000 local profiles evaluated, inconsistencies in structured business data were identified as a statistically significant factor in whether a business was cited or omitted from AI-generated responses.
Generative Citation Gap and Entity Graph Authority
One of the central findings from the study is what Casper Media describes as the generative citation gap -- a measurable disparity in how frequently brands appear in AI-generated responses relative to their actual market presence or search relevance. The gap was found to correlate strongly with entity graph authority, a measure of how thoroughly a brand is represented within structured knowledge systems, including schema markup, third-party data sources, Wikipedia presence, and linked open data.
Enterprise brands with well-developed entity graphs and consistent structured data were cited at substantially higher rates across all three generative platforms. Brands without this infrastructure -- despite having strong traditional SEO signals -- showed citation rates that did not reflect their market position. The finding suggests that optimization strategies built for conventional web search do not translate directly to Generative Engine Optimization, where platforms rely on structured knowledge rather than link-based ranking signals.
The study also found meaningful differences in how ChatGPT Search, Perplexity, and Google AI Overviews source and attribute brands. Each platform weighted entity signals differently, and no single data structure guaranteed visibility across all three, underscoring the need for a multi-model approach to Generative Engine Optimization.
Structured Knowledge Graphs as a Citation Driver
Casper Media's analysis found that brands with formally structured knowledge graphs -- those integrating markup, verified entity records, and consistent data across platforms such as Google Business Profile, Wikidata, and major data aggregators -- demonstrated significantly stronger citation consistency. In contrast, brands relying primarily on traditional content volume or backlink profiles showed higher variability in citation outcomes, particularly on Perplexity and ChatGPT Search.
The implications for enterprise marketing and local business strategy are direct. Citation in AI-generated responses increasingly functions as a new category of brand visibility, distinct from page-one search rankings. Brands not appearing in generative responses risk being absent from a growing portion of information-seeking behavior, particularly as AI-mediated search continues to expand across consumer and professional use cases.
"Entity authority is no longer a secondary technical consideration -- it is a primary driver of how brands appear across generative AI platforms," said Bohdan Hovdia, Founder of Casper Media. "Our benchmark data shows that structured knowledge graphs and multi-model optimization strategies are what separate brands that get cited from those that get overlooked, regardless of their traditional search footprint."
About Casper Media
Casper Media is a digital strategy firm based in Irvine, CA, specializing in AI search visibility, entity optimization, and structured data strategy for enterprise and local business clients. Casper Media LLC conducts applied research and advisory services focused on how brands are represented and cited across generative AI platforms.
Learn more at Casper Media



Tweet
Share