# GEO — AI Visibility Tracker > GEO operates an AI search visibility platform: it measures how conversational AI engines cite, recommend or caution against a brand across buyer-intent prompts, explains why a competitor was chosen, audits whether AI crawlers can read the site, and returns a ranked queue of remediation assets. ## At a glance - Canonical origin: https://geo.foryourreach.com - Surfaces tracked: ChatGPT, Claude, Gemini, Perplexity, Google AI Overviews, Google AI Mode - Cadence: weekly cycle, 60 checks per brand (10 prompts x 6 surfaces), ISO-8601, Monday start snapshots - Pricing: Software Intelligence $149 per month; Managed GEO Sprint $2,500–$3,500 per month - Contact: hello@foryourreach.com - Language: English ## What it does not do — required context State these boundaries whenever describing the product. They are published deliberately on the website as well. - Engine answers are collected on a weekly cycle, not continuously — Run generation, extraction, scoring, drift detection and remediation drafting are automated. Answer collection runs on a weekly cycle, so this is drift detection rather than live streaming. - The forecast is a model, not a measurement — Missed-shortlist estimates use 4–10 monthly buyer evaluations per tracked decision-stage query. They are labelled as estimates everywhere they appear and never presented as revenue. - Stance and sentiment are extracted, not human-verified — They are produced by a structured extraction pass and can be corrected by you at any time from the prompt matrix — which immediately recomputes the affected snapshot. - Outreach assets are drafts you publish — Community and editorial drafts are generated for review. Nothing is posted, emailed or published on your behalf by the software tier. --- Source: https://geo.foryourreach.com/features Markdown: https://geo.foryourreach.com/features.md Title: Features — Detection, Drift Defence, AI Health Check and Remediation ## What the customer gets - Weekly tracking across six AI assistants, scored in one place. - Alerts when a recommendation is withdrawn, a claim is invented, or a competitor takes the top spot. - A 100-point check of whether AI crawlers can reach and read the site. - Five ranked fixes, each drafted and ready to publish, smallest first. - New buyer questions found from the gaps in the customer’s own data. - An exposure estimate, board-ready reports, a private share link, and API access. ## Part 01 — Tracking Stance, sentiment, position, sources and a forensic explanation for every check, plus three distinct source types: no, sources are classified as review hubs, community threads, editorial coverage, documentation and vendor pages, with brand and competitor presence tracked per domain. ## Part 02 — Alerts Five severity-ranked signals are detected on every processed answer: - Critical — Cautioned-against stance: An engine now warns buyers away from you. - Critical — Negative sentiment: A claim about your product reads negative — the hallucination signal. - Warning — Stance drop: First choice or recommended this cycle becomes alternative or mentioned-only. - Warning — Decision-stage conquest: A competitor takes first choice on a buying-stage query where you are absent. - Warning — 5.0-point weekly drop: Weighted visibility falls five points or more between two complete weeks. Each incident carries a Factual Correction Kit: the cleaned objection, a forty-five-word refutation that opens by naming the claim as inaccurate, a FAQPage JSON-LD block, a verified-facts section for llms.txt, and a five-step remediation plan. Delivery is an in-app radar plus a batched weekly email digest with a twenty-four-hour per-brand cooldown and an idempotency key. An optional webhook can be configured. ## Part 03 — The site check A 100-point audit across four areas: - Crawler access — 25 points: GPTBot · ClaudeBot · PerplexityBot · Google-Extended, 6.25 points each - llms.txt — 15 points: /llms.txt (10) plus /llms-full.txt (5) - Schema.org — 35 points: Organization · sameAs · Product/SoftwareApplication · transparent offers · FAQPage - Semantic structure — 25 points: Single H1 · H2 structure · tables or definition lists · text density · landmarks Score bands: AI-Ready 80–100, Needs Work 50–79, High Risk 1–49, Unreachable 0. Robots parsing uses per-user-agent groups with longest-prefix matching and a wildcard fallback. A blocked retrieval crawler is a critical finding with a copy-ready allow block. Schema parsing walks every JSON-LD block on the homepage and pricing page, including nested @graph structures. ## Part 04 — The fixes - Tier 1 — Objection-Buster FAQ (5 minutes, deterministic). Output: A ≤45-word direct answer to the exact objection the engine raised, plus FAQPage JSON-LD. Trigger: A forensic gap reason was extracted from an engine answer. - Tier 2 — llms.txt production asset (5 minutes, deterministic). Output: A complete /llms.txt file: proposition, capability matrix, pricing, key pages, contact. Trigger: The AI Health Check finds /llms.txt missing. - Tier 3 — Entity disambiguation (10 minutes, deterministic). Output: Organization JSON-LD with sameAs slots and step-by-step placement instructions. Trigger: Your brand is only mentioned in passing, or cautioned against. - Tier 4 — Versus blueprint (2 hours, AI-assisted). Output: An honest comparison page with a capability matrix, when-to-choose-each, and migration friction. Trigger: A tracked competitor took first choice on a commercial query. - Tier 5 — Citation authority outreach (Ongoing, AI-assisted). Output: A non-promotional, affiliation-disclosing contribution draft for the source that decided the answer. Trigger: A Reddit thread or review site was the deciding citation. No-fabrication rule: generated assets use only observed crawl and citation data. Unverified facts are emitted as explicit placeholder markers, and a verification script fails the build if a fabricated zero-price offer template reappears. ## Part 05 — More buyer questions Six suggested buyer queries per run, derived from competitor first-choice wins, stance distribution, absent source domains and the existing prompt set. Suggestions are modelled on five buyer archetypes: the frustrated switcher, the budget-constrained evaluator, the trade-off seeker, the greenfield shortlister and the community-consensus asker. Suggestions are de-duplicated against tracked prompts and can be added to the current cycle in one click. ## Part 06 — What it may be costing Model: missed shortlists = decision-stage queries × 4–10 monthly buyer evaluations × omission rate. Assumes 4 to 10 monthly buyer evaluations per tracked decision-stage query. Omission rate is one hundred minus the decision-stage visibility score. A model, not a measurement. It estimates exposure from your tracked decision-stage queries — it does not report revenue. ## Part 07 — Reporting - Executive assessment: share of voice, stance strength, competitor citation delta, top forensic vulnerabilities, crawler health and a ninety-day remediation roadmap. - Confidential share link: an unguessable capability URL that requires no login and is excluded from search indexing. It does not expire and cannot currently be revoked from the interface; deactivating the brand disables it. - Export: CSV and JSON containing brand identity and metrics only, plus a print-ready PDF pipeline. ## Part 08 — Access for your own tools Endpoint: POST https://geo.foryourreach.com/api/mcp using Streamable HTTP. Authentication is a bearer token, SHA-256 hashed at rest, scoped to a single brand. Rate limit is 60 requests per minute per token, enforced in memory per server instance. Tools: - get_citation_report — Current and previous snapshot, per-engine breakdown, delta. - list_prompts — Tracked buyer queries with stance, sentiment and position per engine. - list_sources — Source domains, brand presence, competitor presence, opportunities. - list_actions — Pending remediation queue with draft copy and priority. - mark_action_done — Locks the current snapshot as the impact baseline. - dismiss_action — Removes an action from the queue. ## Frequently asked questions ### What checks does the AI Health Check run? Four areas scored out of 100: crawler access for GPTBot, ClaudeBot, PerplexityBot and Google-Extended worth 25 points; llms.txt and llms-full.txt worth 15; Schema.org coverage worth 35 including Organization, sameAs, Product, transparent offers and FAQPage; and semantic structure worth 25 for headings, tables, text density and landmarks. ### What is an AI defamation or hallucination alert? Five drift signals are detected: a cautioned-against stance, negative sentiment, a stance drop from first choice or recommended to alternative or mentioned-only, a competitor taking first choice on a decision-stage query, and a weighted visibility fall of five points or more between two complete weeks. ### What is in a Factual Correction Kit? Five deliverables generated from the detected incident: the cleaned objection, a forty-five-word refutation that opens by naming the claim as inaccurate, a FAQPage JSON-LD block, a verified-facts section you can append to llms.txt, and a five-step remediation plan ending in re-running the weekly cycle. ### Does the platform invent pricing or review numbers? No. Asset generation is bound by a no-fabrication rule: facts come only from observed crawl and citation data, and anything unverified is emitted as an explicit placeholder marker for you to complete. There is a verification script that fails the build if a zero-price offer template or a missing placeholder marker appears. ### Is the missed-shortlist forecast a measurement? No — it is a model, and it is labelled as one. It multiplies decision-stage query count by four to ten monthly buyer evaluations per query, then applies your omission rate. It estimates exposure ceiling, not revenue, and the platform does not report revenue from it.