Sentryhearth

Founders want to know whether AI answer products mention their company for category questions and whether visitors arrive from AI referral links. Sentryhearth runs a declared sample of prompts under permitted provider access, records mentions, citations and co-mentions and combines them with a privacy-minimized site referral signal. The supplied research confirms several visibility trackers and a separate traffic-attribution product; none in the review combined both. The market is therefore active and the low-end gap narrow. Model answers are stochastic, personalized, region- and version-dependent. A prompt sample is not buyer demand, ranking is not search rank, sentiment is an unreliable inference and mention share is not market share. Referral data is also incomplete because browsers, applications and privacy controls strip source fields; traffic labeled direct cannot be reassigned with certainty. Prompt observation, model version, mention, citation, referral candidate, session event, conversion, revenue and business outcome remain separate. Collection must honor provider and site policies and avoid personal profiling. Success is a reproducible historical sample and better source visibility—not proof of recommendation, discoverability, influence or revenue attribution.

A solo founder or small growth team monitoring a narrow product category and owning a privacy-conscious website analytics setup.