We Don't Reverse-Engineer the Algorithm. We Instrument It.
When a buyer asks an AI assistant for a local business, the assistant rewrites the prompt into hidden retrieval sub-queries, fetches web pages, and cites what it fetched. Measure the exact sub-queries you need to win and lock your exclusive territory seat.
Is ChatGPT, Claude, or Perplexity Recommending Your Business?
No invented scores. See the exact buyer questions tested, how many answers cited your domain, which competitors were cited instead, and the exact retrieval sub-queries to win.
1. Measure Sub-Queries (Harvester): Every Monday, we ask ChatGPT, Claude, and Perplexity up to 20 buyer-intent questions (plus real Google Search Console queries with 10+ impressions) and record the exact sub-queries they run.
2. Publish Verified Answer Blocks (RLO-2 Pages): For every sub-query you are missing, we build a direct 50-word sourced answer block with FAQPage & LocalBusiness schema and submit it via IndexNow & Search Console.
3. Track Week-Over-Week Citation Share: Append-only weekly windows (`2026-W42`, `2026-W43`, `2026-W44`) prove whether citations moved. Expected timelines: first AI citations in 4–8 weeks, Google ranking movement in 3–6 months.
TERR-1 Territory Seats — Capped by Measured Citation Capacity
Default capacity is 5 seats per trade per city, plus 1 neighbourhood exclusive. Once 4 weeks of RLO probe data exist, seat capacity is capped at `min(5, measuredCapacity)` — so we never sell more seats than AI assistants actually cite in your market.