Context
CITECHOICE uses controlled replay of authentic multi-turn agentic-search transcripts to test how document presentation and order affect citation allocation when competing sources support the same fact.
What changed
Structured rendering increased target citation count by 0.50 citations per answer in the main controlled replay, while the effect on whether the target was cited at all was inconclusive. Fresh decoding also exposed substantial decision noise.
Why it matters
AI visibility measurement should distinguish retrieval or source admission from downstream citation allocation, and should account for stochastic citation decisions rather than treating one observed citation as stable.
Visibility OS interpretation
Within already retrieved frozen transcripts, presentation can causally redistribute visible citation credit. This does not establish that structured formatting reliably causes retrieval, source admission, or higher ranking in production systems.
What remains uncertain
How well do these effects replicate across named production answer engines, languages, query classes, and live retrieval pipelines? Which structural changes affect admission versus only citation allocation?