Vercite Research

We measure how AI actually answers.

Original studies built on the Vercite platform’s own measurement data: real prompts, real responses, real citations, at sample sizes that mean something. We publish what the engines do, not what they claim.

The evidence base
5.31MCitations analysed
158,847Domains classified
50,307Query fan-outs analysed
5AI engines compared

Aggregated across the published studies below. Every figure measured by the Vercite platform, not estimated or surveyed.

How we work

Platform data, not surveys

Every number comes from the Vercite measurement pipeline: real prompts, real engine responses, real citations, captured as they happen. Nothing is self-reported and nothing is sampled by hand.

Scale before stories

A pattern only gets published when it holds across the full dataset: tens of thousands of data points, not a handful of screenshotted chats.

Like-for-like comparisons

The same prompt sets run across every engine, side by side. When a chart shows a difference, it is a difference in the engine, not in the question we asked it.

Try Vercite for free today

Monitor your brand across ChatGPT, Gemini, Perplexity & Google AI

Get started free