Loading…
Why publish this. SEO agencies sell ranked-list traffic. GEO is different: it’s the citation pattern inside the assistant’s answer. If a buyer asks an AI for "best ERP consultants 2026," the recommendation set varies a lot by engine. This dashboard makes that visible — for you, for your competitors, and for us.
Methodology. Same query, same day, fired at four engine slots (ChatGPT / Claude / Perplexity / Gemini) — but only the local-model rows (ChatGPT + Claude, one on-device gpt-oss:20b) return data today; Perplexity and Gemini are dark and show polled:0. pollLane.js writes per-engine JSON snapshots. build-engine-benchmark.mjs rolls the last 30 days into /assets/data/engine-benchmark.json. Read each column’s polled count before trusting it — a column at 0 did not run.
What this isn’t. It’s not a ranking authority. Engines disagree, indexes shift week-to-week, and any single-day snapshot has noise. The 30-day rolling smooths that. For a deeper read on a specific category, the paid GEO subscription runs 12 queries × 4 engines × 30 days against your domain plus competitors with full attribution.
See the GEO subscription → Free 1-query audit