Thursday Sep 3
RESEARCHPerplexity

A Third Of Perplexity's Citations Don't Check Out

2SEP
CITED EMPTY

Haus Research fetched every source Perplexity cited across 310 questions. A third of the links attached to a number would not open or never held that number. The citation looks like proof.

1,826 citations were bolted onto sentences stating a figure. 34.7% pointed at a page an ordinary reader could not open, or a page with none of the figures in the sentence.

Scored per claim instead of per link, 14.4% of 872 numeric claims had no support behind them at all. Same audit, gentler denominator, still bad.

The setup: two Perplexity search models, 210 technology companies, English-language questions, audit run September 2. Every cited URL was actually fetched and read.

full brief & sources

⚡ Why this matters

  • Citations are the whole product promise. Perplexity sells sourced answers, not vibes.
  • The failure mode is invisible. A link that resolves looks verified. Nobody clicks.
  • Every agent you ship that cites its work inherits this exact problem.

🔍 What happened

  • Haus Research asked two Perplexity search models 310 factual questions about 210 technology companies.
  • They collected every cited source, fetched it, and checked whether the page said the thing it was cited for.
  • 34.7% of 1,826 figure-bearing citations failed. Either the page would not open, or it contained none of the figures in the sentence.
  • Per-claim scoring: 14.4% of 872 numeric claims had zero supporting evidence.
  • Audit date: September 2. Scope: English-language questions, technology companies.

💬 Smart takes

  • The audit's own framing is careful. It counts a fail only when the page has none of the numbers, not when a number is merely hard to find.
  • Broader work this year lines up. Six studies covering 366,087 real-world citations found the same pattern of misattribution across engines.
  • Skeptics will note the domain is narrow. Company financials are exactly where numbers move fastest and stale pages are most likely.

🧭 Where this goes

  1. Likelycompetitors ship citation verification as a feature. 'We fetch and check every link' becomes a marketing line.
  2. Possiblean enterprise buyer makes citation accuracy a procurement requirement. That would reset the category.
  3. Wild CardPerplexity publishes its own audit with a different methodology and the numbers become a public fight.

🥄 The Spoon Take

This is the measurement problem, not a Perplexity problem. Any system that attaches a source to a sentence is making a claim it never verifies. If your product cites anything, go fetch your own links and check them. You will not like the number either.

🤔 Pushback

One vendor, one domain, one week. A 34.7% link-level failure rate is not the same as a 34.7% wrong-answer rate, and the report does not claim it is.

·A third of Perplexity's citations don't contain the number they're cited for (Haus Research)·Haus Research - independent research on companies, markets and AI assistants·How AI Engines Cite the Web: The Six 2026 Studies (Everything PR)·Perplexity citation audit (Salo linkspam)