Real-device evidence now. Systematic market Index next.
The current library shows how ChatGPT, Gemini, and Perplexity answer commercial local prompts across captured client markets. The public AI Local Visibility Index will publish after structured market-study sidecars land.
6,946 real-device rank-tracking observations across ChatGPT, Gemini, and Perplexity.
2,892 multi-engine groups in the June 12 audit; 71.1% had different visibility status across engines.
292 pixel-rank verified movements, 127 verified #1 outcomes, 14 days median movement window.
25 frozen metro points are ready. The public market Index publishes after structured sidecar data lands.
Boundary note: this page uses rank-tracking evidence and verified movement records. It is not yet the completed systematic metro x vertical x prompt market Index.
same-query groups where engines disagreed on the #1 across 2,328 same-query comparisons. Rank-tracking evidence library, not the completed systematic metro Index.
One engine is not an AI strategy.
In the current rank-tracking evidence library, side-by-side engine checks repeatedly disagree. The older public rollup shows a 90.7% #1 disagreement rate across 2,328 same-query comparisons; the June 12 bucket audit also found 71.1% different visibility status across 2,892 multi-engine groups.
of the older public rollup's side-by-side comparisons produced a different #1 business depending on which AI you asked.
Same practice, same question (“veneers dentist,” Los Angeles), same day. ChatGPT seats it at #4; Gemini buries it at #12; Perplexity puts it last of four. Only a screenshot proves which customers you are losing.
Gemini gives out #1 the most.
Share of rank-tracking observations where the engine named a single business as the top recommendation. The engine your customers use changes your odds dramatically.
Rank in AI answers moves.
Across the businesses our capture network measures, 225 climbs into the top-2 passed the pixel-verification gate - each documented by before/after captures from the same device asking the same query (Jun 10, 2026 snapshot).
before#14Apr 28, 2026Open full
after#1 of 15May 12, 2026Open fullTNT Design & Build - Rancho Santa Fe, CA - “remodeling companies”
before#25Apr 28, 2026Open full
after#2 of 4May 12, 2026Open fullBizWell Benefits - Texas - “self-employed health insurance”
The systematic market Index is next.
Screenshots are evidence. Structured sidecars are the dataset. The public Index publishes when both the receipt and the parsed answer record are present.
We already have the metro anchors staged. The public Index will wait for index/date captures and parsed sidecars with businesses named, citations, accuracy flags, and map-pack comparisons.
Want the proof layer for your market?
The paid baseline turns the evidence method into your competitor map, category map, reputation ledger, and priority-weighted execution plan. We capture the current answer, identify why AI names competitors, and re-measure the same market prompts in 14-30 days.