> ## Documentation Index > Fetch the complete documentation index at: https://docs.getcited.in/llms.txt > Use this file to discover all available pages before exploring further. # What we don't do and why > The boundaries of what Cited measures and claims — what's deliberately out of scope, and the reasoning behind each line. Knowing what a product does *not* do is as important as knowing what it does. This page documents the boundaries Cited has drawn around its measurement, so you can interpret your dashboard with the right expectations. ## We don't claim deterministic measurement LLM responses are inherently [non-deterministic](/concepts/foundations/non-determinism) — the same prompt asked twice can produce different brand lists. Cited provides aggregated metrics that are statistically stable at sufficient sample sizes, not exact repeatable readings. This is a property of the measurement domain, not a fixable bug. Any tool that claims perfect repeatability on AI-generated content is over-promising. ## We don't grade your content Cited measures brand visibility in AI answers. It does not grade, score, or evaluate the quality of your own content — blog posts, product pages, documentation. Content quality assessment tools exist; Cited is not one of them. The [GEO Score](/methodology/geo-score) is a separate, narrower product: a technical readiness check on your web presence for AI discoverability (robots.txt configuration, schema markup, llms.txt presence, HTML structure). That's a technical audit, distinct from content quality grading. ## We don't guarantee visibility improvements Cited provides measurement and diagnostics. It does not guarantee that acting on a recommendation will produce a specific increase in [mention rate](/concepts/metrics/mention-rate) or [share of voice](/concepts/metrics/share-of-voice). AI platform behavior is controlled by the platforms themselves — OpenAI, Google, Anthropic, Perplexity, xAI. No external measurement tool can guarantee outcomes in a system it does not control. Cited identifies where opportunities exist and measures whether changes are working, but the actual lift depends on how the AI platforms respond to content and authority signals. ## Impact Scores are about priority, not prediction [Impact Scores](/methodology/impact-scoring) tell you which recommendation to act on first — they don't predict the exact mention-rate or share-of-voice lift you'll see. A score of 60 means "60 out of 100 on this brand's priority scale", not "+60 percentage points if you do this." ## We don't track individual user conversations Cited queries AI platforms programmatically and measures their public-style responses. It does not monitor, intercept, or analyze individual end-user conversations with AI platforms. The data is about brand presence in AI answers, not about user behaviour or identity. ## Related concepts * [Non-determinism](/concepts/foundations/non-determinism) — why deterministic measurement is not possible * [Reading your data with confidence](/methodology/data-freshness) — when to trust a movement * [Impact scoring](/methodology/impact-scoring) — how recommendation priority is calculated ## Frequently asked questions Because measurement credibility depends on honesty. A tool that claims to measure everything perfectly is not trustworthy. Stating boundaries explicitly helps you interpret the dashboard correctly and decide what the data can and cannot tell you. Yes. Planned additions include broader coverage of Google's AI surfaces in cross-platform benchmarking, expanded prompt libraries, and richer citation-source mapping. These are prioritized against customer value and engineering cost.