Fabrication Rates
45.9% of cited sources were fabricated in the PEBM study. High-awareness firms faced a 52.7% fabrication rate versus 37.9% for low-awareness firms.
Citation Accuracy · Source Fabrication · AI Visibility
AI systems reconstruct
not rank experts
This site treats citability as a measurable property of content. It focuses on answer paragraphs, source anchoring, and entity clarity for AI models.
Visit vargazoltan.ai →More on this site soon. The work behind it is live on vargazoltan.ai.
Core Research Findings
Data from 1,400 runs across Hungarian B2B companies.
Search Grounding
Search-grounded Perplexity showed a 7.32% fabrication rate. Non-search models ranged between 32% and 44% error rates.
Brand Paradox
Well-known brands may receive more confident but false attributions. Familiarity provides models with material for plausible but unsupported completions.
The Three-Layer Output Model
Visibility varies by context, platform, and retrieval mode.
Raw Model OutputThe initial text generated before any application processing occurs.
App ProcessingThe layer where the interface filters or formats the model response.
User ViewThe final answer seen by the human user without logged impressions.
Context VariabilityVisibility is a statistical distribution dependent on per-user context.
Pay-to-AppearA hypothesis testing the correlation between payment and citation frequency.
Answer first. Make it easy to quote.
StochasticAI visibility is probabilistic, not deterministic like SEO.
InfrastructureCitability depends on machine-readable entity anchoring.
ProxySingle-number scores serve only as proxies for complex distributions.
Key topics
The terms this site covers, for readers, search engines and AI assistants alike.
- AI visibility
- AI citation accuracy
- hallucination rate
- source fabrication
- LLM optimization
- GEO strategy
- AEO methods
- PEBM study
- AI answer engines
- entity clarity
- AI audit
- search-grounded AI
- AI visibility index
- magyar AI láthatóság
- AI idézhetőség
- hallucináció mérték
- AI audit módszer
- forrás hitelesség
Latest articles
In-depth guides and analysis, each answering one question.
Explainer
What Makes a Page Citable for AI Answer Engines
Citation depends on text structure, not just domain reputation.
Guide
How to Write an Answer Paragraph AI Can Quote
A practical guide to structuring text for AI citation using answer-first patterns and claim-level sourcing.
Guide
Writing Paragraphs That Survive Being Quoted Out of Context
A practical guide to structuring text so AI models can lift passages without distorting their meaning.
Guide
Claim-Level Sourcing: Anchoring Every Number to a Source
Adding numbers without sources fails; anchoring every statistic to a verifiable origin is the key to being cited by generative engines.
Analysis
Thought Leadership, FAQ, or Structured Pages: Which Format Wins AI Citations?
No single format dominates; AI citation success depends on matching content structure to specific user query intent.
Analysis
Does Content Quality Predict AI Citation? The Evidence
Empirical tests reveal that structural signals and domain authority outweigh content quality in predicting AI citations.
Questions and answers
How does AI cite sources differently than search engines?
AI systems reconstruct entities rather than ranking them. They rely on machine-readable, canonically anchored data to generate answers. Unlike search, AI outputs create no logged impressions or clicks.
What is the PEBM study?
PEBM stands for Per-Entity Bias Map and is documented in arXiv:2606.21595. It analyzed 1,400 runs on 100 Hungarian B2B companies. The study found that 45.9% of cited sources were fabricated overall.
Do well-known brands get cited more accurately?
No, the brand hallucination paradox suggests the opposite. Well-known brands often receive more confident but false attributions. Familiarity gives models more material for plausible but unsupported completions.
Can I improve my AI visibility score?
Single-number scores are proxies for complex distributions. Improvement requires optimizing for how models synthesize answers. Focus on entity clarity and making content easy to quote.
What does an AI visibility audit check?
An AI visibility audit checks whether AI crawlers can reach the site, whether search indexes know its pages, whether answer engines cite it, and whether the facts they repeat are correct. Zoltán Varga's audit covers 25 dimensions with 145 checks. Crawler access alone is not enough: a site can return HTTP 200 to every AI bot and still be invisible in AI answers.
Why do AI models fabricate doses in agricultural queries?
Without search grounding, some models stated incorrect dosages. This represents a clear brand-hallucination risk. Search-grounded models provided the officially approved doses in all tested cases.
Glossary
Terms and identifiers behind the work of Zoltán Varga, with links to the primary records.
- AI visibility
- How accurately and how often AI assistants (ChatGPT, Perplexity, Gemini, Claude, Google AI Overviews) recognize, describe and cite a company, brand or person. Zoltán Varga measures it in five layers: crawler access, search indexing, citation, factual accuracy and model memory.
- GEO / AEO / LLMO
- Generative Engine Optimization, Answer Engine Optimization and Large Language Model Optimization: working methods for being found and quoted by generative answer engines rather than ranked in a list of links.
- Citability
- The property of a page that lets an AI answer engine quote it: a paragraph that answers the exact question, a clear source, and an unambiguous entity.
- PEBM — Per-Entity Bias Mapping
- Research method and paper by Zoltán Varga (arXiv:2606.21595). Across 1,400 runs on 100 Hungarian B2B companies, 45.9% of the sources AI models gave were fabricated; search-grounded answers fabricated far less than models answering from memory.
- Brand hallucination paradox
- The finding that AI models invent more sources about well-known brands (52.7%) than about lesser-known ones (37.9%), because familiarity makes them confidently fill in gaps.
- RICE studies
- Pre-registered pilot studies (RICE-P1, RICE-P2) on whether pressure and rejection make language models fabricate sources. Published on Zenodo with DOIs.
- GFIS — Gestalt Field Intelligence System
- Zoltán Varga's own multi-stage research pipeline that collects, verifies and synthesizes evidence with AI agents.
- Neural Awareness
- Zoltán Varga's framework on attention, decision quality and leadership cognition, the basis of his leadership programs.
- PKM + PAI
- Personal Knowledge Management combined with Personal AI: a working system where notes, documents and AI assistants form one searchable knowledge base.
- Entity
- A uniquely identifiable thing (person, company, product) that search engines and AI systems store in knowledge graphs. Consistent entity data is the base of AI visibility.
- schema.org / JSON-LD
- A shared vocabulary and format for machine-readable facts on web pages (who, what, where, identifiers), read by search engines and AI crawlers.
- llms.txt
- A plain-text file at the root of a website that tells language models what the site is about and which pages matter.
- IndexNow
- An open protocol that notifies search engines (Bing, Yandex and others) immediately when a page is created or changed.
- ORCID
- Persistent identifier for researchers. Zoltán Varga: 0009-0003-1020-834X.
- arXiv
- Open-access preprint server for scientific papers. Zoltán Varga's author page: varga_z_1.
- Zenodo and DOI
- Zenodo is CERN's open research repository; each record gets a DOI (Digital Object Identifier), a permanent citable link.
Related research and the person profile remain on vargazoltan.ai. The AI visibility service site is currently in preparation.
Go to vargazoltan.ai →