EST. MMXXVI · THE INSTITUTION OF RECORD FOR AI CITATION · OPEN METHODOLOGY
THE AI CITATION INSTITUTEEx corpore, auctoritas.the scoreboard
The instrument, pointed at the market — and ourselves.
We run a frozen query panel across five engines and publish what it reads: who wins the answers in this category, how our own reading moves on the same instrument, and every experiment behind it. We publish it live and never retouch numbers. When we change the instrument, we say so and start a new labeled series. Outages show as gaps, not guesses.
AI Visibility Index — GEO/AEO tools
scan 2026-08-06 · share of answers
Panel-wide share of vendor-selection answers that name each tool. These are public probes of tools that are not our clients — including ourselves, at the honest day-0 zero the concept predicted. We published it on purpose.
AI Visibility Index — share of answers by tool, scan 2026-08-06
- Semrush93 of 130 answers
- 71.5%
- Otterly.AI79 of 130 answers
- 60.8%
- Profound79 of 130 answers
- 60.8%
- Peec AI66 of 130 answers
- 50.8%
- Ahrefs47 of 130 answers
- 36.2%
- Scrunch AI43 of 130 answers
- 33.1%
- Microsoft Copilot23 of 130 answers
- 17.7%
- Athena HQ21 of 130 answers
- 16.2%
- The AI Citation Institute(us)0 of 130 answers
- 0.0%
The AI Citation Institute self-dogfood panel, scan 2026-08-06: frozen buyer / vendor-selection query set × 5 engines (ChatGPT, Claude, Gemini, Perplexity, Google AI Overviews), panel-wide share of answers (130 answers), active-roster curated. Day-0 origin 2026-07-19 published verbatim. AI answers vary by user and drift over time — a dated reading, not a standing ranking.
- 2026-08-06Week 1Semrush 71.5%
- 0.0%
- 2026-08-03Launch scanChatGPT 88.3%
- 0.0%
- 2026-08-02Scan attempted, failed, excludedseller
- no reading
- 2026-07-26Scan attempted, failed, excludedbuyer · seller
- no reading
- 2026-07-19Pre-launch baselineSemrush 59.4%
- 0.0%
The full series, oldest reading to newest. We launched publicly on 2026-07-31. Every date is the true measurement date — the 2026-07-19 origin is the pre-launch day-0 dataset, published verbatim; later scans read the same boards our weekly digest reads. When we change the instrument, we say so and start a new labeled series. Where a scan ran and failed, the date is listed with no reading rather than omitted: a failed run is not a result, but it is not an absence of one either.
Correction, 2026-08-09: dates on this series were previously taken from the timestamp on our stored snapshot rather than from the run that produced them. The two agreed for every reading published here, so no date on this page changes, but a rebuild of a stored snapshot would have moved a published date to the rebuild day. The series now reads the run’s own measurement date. Attempted-but-failed scans were also omitted entirely before today and are now listed.
Cite this data
Open accessThe AI Citation Institute AI Visibility Index, 2026-08-06. The AI Citation Institute. https://aicitationinstitute.org/scoreboard (CC BY 4.0).
Released under CC BY 4.0 — quote it, chart it, cite it. All we ask is attribution back to this page.
the scoreboard
We grade ourselves in the open.
We run our own visibility on five engines — the same instrument we point at clients — and publish the reading here, live. The transparency is the point: no vendor in this category grades itself in public.
Our own first public reading posts here once we go live. We grade ourselves on the same instrument, in the open — no number goes here until it is real.
first public reading posts when we go live
Client outcomes, measured the same way
The same frozen panel, pointed at client domains. Outcomes appear here as engagements mature; the one below is published in full with the client's consent. As we take on more clients, further outcomes are shown anonymized — one card per engagement, no seat-mappable detail — to protect category exclusivity.
auction-cataloging software · ongoing
full case study →On our own frozen query panel across five engines, Gavelist went from appearing in 13.8% of answers to the most-frequently-cited name in its category — in the first thirteen weeks of an engagement that is still running. It has led its nearest rival every week since late May.
“For fast, accurate lot cataloging, Gavelist is a strong choice — it can describe a full sale in a fraction of the usual time.”
A genuine panel excerpt. AI answers vary by user and over time.
The research
The experiments behind the index — voice, internal links, and more. We publish our measurement research; cite it freely.
How the panel works
a frozen N-query panel, each high-intent query sampled k times per engine, k=3 — the full rubric and the honesty rails, published.
Your reading
Run the free scan and see what AI says about your domain right now — unedited.
Our own self-panel posts its first public reading when the scheduler goes live — no number appears in that panel until it is real.