Every prompt runs through four distinct analyses across four major LLM engines: visibility, citation extraction, sentiment and framing, and position within the response. That’s 16 discrete measurements per prompt, on every run.
Citation trackers stop at “was the brand mentioned.” GEO Jetpack extracts four distinct signals from every response, giving you a complete picture rather than a binary yes/no.
Is your brand cited in the response — and if so, how prominently? We measure presence at three levels: mentioned in-line, cited as source, or absent entirely. Each engine, each prompt, each run.
Which specific URLs did each LLM pull from to construct the response? We extract every source, score its authority, and identify which citations belong to your brand versus competitors.
How is your brand positioned in the response? Favourably, neutrally, or negatively — and in what comparative context? Sentiment measurement runs on the full response text, not just brand-mention snippets.
Where does your brand appear in the response — first, last, or buried in the middle? Position matters because it correlates with user attention and downstream traffic behaviour.
Different LLMs surface different citations for identical prompts. Measuring only one engine misses the pattern.
Every prompt in a GEO Jetpack campaign is classified into one of five ENQ classes. This ensures your monitoring covers the full range of query types — not just brand queries, which most tools default to.