A short primer below. The full methodology page — with worked examples, the research behind the taxonomy, and the reasoning behind each class — is being published shortly.
ENQ™ — Entity-based Natural Query classification — is the taxonomy underneath every campaign GEO Jetpack runs. Rather than treating all prompts as equivalent, ENQ™ classifies each into one of five structural classes, so measurement reflects the shape of the queries being asked, not just their volume.
Most AI Search monitoring tools count prompts equally: 100 prompts, 100 data points. But a brand-direct question (“Tell me about Aviva”) sits in a completely different competitive landscape from a topic-discovery question (“How does pension consolidation work”). Treating both as one aggregate visibility score obscures where a brand is actually winning or losing.
ENQ™ separates prompts into five classes so movement in one class isn’t hidden by noise in another. A brand can be winning brand-direct queries decisively while quietly losing topic-discovery queries — ENQ™ makes that visible, and prioritises the recommendation accordingly.
Every campaign GEO Jetpack runs distributes prompts across the five ENQ™ classes proportionally, so no single query type dominates the monitoring signal. Reports break out visibility, citation, and sentiment by ENQ™ class — meaning you see not just how your brand is performing in AI Search, but where the performance is concentrated.
Recommendations follow the same logic. If Class B (competitive) visibility is soft while Class A (brand-direct) is strong, the fix is different from the reverse case. ENQ™ is what allows the recommendations to be specific rather than generic.
The full page will include the research programme behind the classification (RAIDAR), worked examples across three industries, the paper on why structurally-distinct prompts outperform paraphrased variants for citation discovery, and the technical detail of how classes are assigned at ingestion.
If you’d like a heads-up when it’s published, email us and we’ll send you the link. If you have specific questions about the methodology in the meantime, email Joe directly.