Study ties AI citation likelihood to how recently a page was updated
Corrections
- 2026-11-07: The 72% versus 42% comparison classifies the same 4,124 cited pages by update date and publication date. It does not show that refreshed old pages outperform newly published ones. The answer and interpretation are corrected.
What this answers
- How much does content recency affect AI search citations?
- A Seer Interactive analysis of 47,097 citations across ChatGPT, Gemini and Perplexity found 75% of cited pages had been updated within the past year and 88% within two years.
- Does refreshing an old page work better than publishing a new one for AI visibility?
- This study cannot establish that. It classified the same 4,124 cited pages by two dates: 72% were less than a year old by last update, versus 42% by initial publication. There was no comparison of citation probabilities for refreshed and newly published pages.
Seer Interactive analyzed 47,097 citations across ChatGPT, Gemini and Perplexity for four brands over a four-month window and found 75% of cited pages had been updated within the past year, rising to 88% within two years. Among 4,124 pages with both dates available, 72% looked fresh by last-modified date versus only 42% by original publish date, and the observed recency shares varied by engine — Gemini 78%, ChatGPT 73%, Perplexity 65%.
The update-date comparison uses the same cited pages, not competing cohorts of refreshed and newly published pages; it does not establish a causal benefit from refreshing.
Why it matters: The study describes the dates on cited pages. Its denominator and lack of an uncited comparison group limit what it can establish about a refresh strategy.
The record: ChatGPTGeminiPerplexity
Posted to the wire August 6, 2026. Updated November 6, 2026. Edited by Joe Balewski.