# What is Prompt-to-citation tracking?

Prompt-to-citation tracking is the practice of measuring AI visibility by running a fixed set of buyer prompts across AI engines on a cadence and recording which brands get cited, since standard web analytics cannot see most of it.

## How it works

You build a representative set of prompts, run them across the target engines repeatedly, and log the citations. This produces a measurable read on whether you appear in AI answers and how that changes over time.

It exists because AI traffic is largely invisible to default analytics. In rawmktg's testing, [GA4 misses roughly 30% of AI referrers and misclassifies many of the rest](/blogs/prompt-to-citation-tracking), so a prompt-based method is the only reliable way to see AI presence.

## Prompt-to-citation tracking vs rank tracking

Rank tracking records where your pages sit in classic search results. Prompt-to-citation tracking records whether your brand is named in AI answers. The first measures position in a list; the second measures presence in the synthesised response, which is where AI-era visibility actually lives.

## Why it matters for B2B

It is how GEO becomes accountable. Without it, AI visibility is a guess; with it, you have a number you can benchmark, tie to pipeline, and report to the business.

**Common mistake**

Trying to read AI visibility out of GA4 alone. It misses roughly 30% of AI referrers and misclassifies many of the rest, so the only reliable read is to run a fixed prompt set against the engines and record the citations yourself.

*Source: https://rawmktg.com/glossary/prompt-to-citation-tracking · rawmktg. by Vinayak Ravi*