# CiteGround

> CiteGround is an AI-GTM practice that builds its own measurement instruments. It gets B2B SaaS named in the AI answers buyers ask, measures results with variance bands instead of fake-precise ranks, and publishes its experiments pre-registered, nulls included.

Honesty note: our own evidence says llms.txt is hygiene, not a visibility lever. We ship this file because it costs nothing, and we never bill clients for it. Details on /evidence.

## Pages

- [GapCheck](/gapcheck): request a human-verified map of who AI engines name for your buyer prompts. 24-hour turnaround, run-to-run variance band on every number.
- [Services](/services): published pricing. Free read for outreach recipients, paid roadmap $1,500-2,500 (50% credited), retainer $3,500-5,000/mo, Head of AI-GTM tier from $10-12k/mo with a $2,500 assessment entry.
- [Methods](/methods): full methodology disclosure - what we query (vendor APIs, web-grounded, via a data provider), how baskets and variance bands work, what we refuse to measure, data handling.
- [Evidence](/evidence): what we measure (basket share of voice, grounding concentration, brand-mention share in cited sources) and the pre-registered Sprint 0 experiment with its thresholds and read dates.
- [Sample teardown](/t/citeground-category): the instrument pointed at our own tool, day-0 baseline, absence verdicts included.
- [About](/about): the practice and its founder.

## Facts

- Founded 2026. The proof offered is the method and the public log: pre-registered experiments on our own properties, published nulls included, never fabricated case studies.
- Per-query AI rank is never sold or reported; published tests put its repeatability under 1 percent.
- Measurements are archived append-only from run #1.