Free original research

The measured state of AI search

Name: Clear Cited AI Visibility Index
Creator: Clear Cited
License: https://creativecommons.org/licenses/by/4.0/

By Logan E Adams, Founder Reviewed & updated Sep 2026 Measurement-first: figures are median share-of-model with 95% confidence intervals. How we measure.

We publish what ChatGPT, Perplexity, Claude, Gemini, and Grok actually recommend — ranked by share of model, measured reproducibly, with the data and method open.

Measured: at least 10 runs per engine, spent adaptively, across 5 AI engines · 95% confidence interval on every figure · open data (DOI 10.5281/zenodo.21612952 (opens in a new tab)) · how we measure

Read: State of AI Search Browse the Index

Reports & datasets

Report

State of AI Search

The headline finding: engines disagree on the “best tool,” and ~95% of the citations behind their answers are third-party. What that means for AI visibility.

Leaderboards

AI Visibility Index

Per-category share-of-model leaderboards (CI/CD, observability, and more), refreshed monthly. Datasets published CC BY 4.0.

Living study

Do directory listings change AI citation?

A pre-registered, living before/after measurement on ourselves — does buying directory listings actually change whether AI engines cite you? Null results published, method open to challenge, raw data ungated.

Coverage

Cited by

Every third-party page that cites or covers this research, with dates and links. Published even when it is empty.

Method

How we measure

Every buyer prompt runs 10+ times per engine; we report the median share of model with a 95% confidence interval. Consumer apps and APIs differ — we say so.

What we measure right now

Share-of-model leaders across the categories in our AI Visibility Index. Measured rows are medians over 10+ runs per engine; share-of-model carries a Wilson 95% confidence interval over the pooled mention denominator (presence a Wilson 95% CI). Illustrative previews are clearly marked and are not live measurements. Coverage: 9 measured, 0 illustrative.

AI Visibility Index share-of-model leaders by category, with 95% confidence interval and measurement status.
Category	Leader	Share-of-model	95% CI	Status
AI observability tools	Datadog	18.2%	16.8–19.8	Measured 2026-08-07
API platforms	Kong	21.2%	19.4–23.1	Measured 2026-08-07
CI/CD platforms	GitHub Actions	21.5%	19.7–23.3	Measured 2026-08-07
CRM software	HubSpot	23.2%	21.4–25.0	Measured 2026-08-07
Databases	PostgreSQL	21.7%	20.0–23.5	Measured 2026-08-07
Feature flag platforms	Flagsmith	16.8%	15.3–18.4	Measured 2026-08-07
Incident management platforms	PagerDuty	24.6%	22.6–26.7	Measured 2026-08-07
Product analytics platforms	Mixpanel	24.0%	22.1–25.9	Measured 2026-08-07
Vector databases	Qdrant	19.7%	18.1–21.4	Measured 2026-08-07

Browse every category and download the datasets (CC BY 4.0) on the AI Visibility Index. AI answers are non-deterministic; figures reflect a point in time and will drift. We do not guarantee any ranking or citation.

Honesty. A single AI screenshot is one sample from a distribution. We never publish a figure we have not measured this way, and figures marked as a methodology preview are illustrative until the first real public run.

The research programme

A research programme, not a blog: every study is pre-registered before we look, ships with a limitations block, and is designed to be repeated — the single biggest structural gap in a landscape of one-off vendor snapshots. Publication doesn't depend on the direction of the result; a null publishes with the same production values as a win.

The research registry: every study, its tier, status, pre-registration, DOI, the SKU it evidences and how to cite it.
Study	Tier	Status	Pre-registration	DOI	Evidences	Cite this
AI Answer Volatility Index	Tier 1	running	sealed in-repo, public registration pending — the hash chain is attested by us, which is worth what your trust in us is worth	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	measured, growth_retainer…	Cite `Logan E Adams (2026). AI Answer Volatility Index. Clear Cited. https://clearcited.com/research/s-volatility/`
Cross-Engine Disagreement Index	Tier 1	proposed	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	measured, full_audit…	Cite `Logan E Adams (2026). Cross-Engine Disagreement Index. Clear Cited. https://clearcited.com/research/s-disagreement/`
Newsletter/email presence vs AI citation	Tier 1	proposed	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	growth_retainer, scale_retainer…	Cite `Logan E Adams (2026). Newsletter/email presence vs AI citation. Clear Cited. https://clearcited.com/research/s-newsletter-citation/`
Schema markup on NEVER-cited pages	Tier 1	proposed	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	growth_retainer, addon_technical_seo	Cite `Logan E Adams (2026). Schema markup on NEVER-cited pages. Clear Cited. https://clearcited.com/research/s-schema-never-cited/`
Content-freshness refresh test (causal, matched control)	Tier 1	proposed	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	addon_blog_post, growth_retainer…	Cite `Logan E Adams (2026). Content-freshness refresh test (causal, matched control). Clear Cited. https://clearcited.com/research/s-freshness-refresh/`
Developer-doc / technical-content citation study	Tier 1	proposed	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	growth_retainer, scale_retainer…	Cite `Logan E Adams (2026). Developer-doc / technical-content citation study. Clear Cited. https://clearcited.com/research/s-devdocs-citation/`
GEO industry citation-hygiene audit	Tier 1	proposed	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	house authority	Cite `Logan E Adams (2026). GEO industry citation-hygiene audit. Clear Cited. https://clearcited.com/research/s-citation-hygiene/`
The directory study (does buying listings change AI citation?)	Tier 1	running	sealed in-repo, public registration pending — the hash chain is attested by us, which is worth what your trust in us is worth	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	addon_directory_boost, addon_directory_growth…	Cite `Logan E Adams (2026). The directory study (does buying listings change AI citation?). Clear Cited. https://clearcited.com/research/s-directory/`
The negative-results ledger (what didn't work)	Tier 1	running	sealed in-repo, public registration pending — the hash chain is attested by us, which is worth what your trust in us is worth	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	measured, growth_retainer…	Cite `Logan E Adams (2026). The negative-results ledger (what didn't work). Clear Cited. https://clearcited.com/research/s-negative-results/`
AI citation -> funnel/revenue	Tier 2	proposed — blocked: Reach 3 clients with sealed T0 baselines — 3 studies cannot commence below it	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	measured, growth_retainer…	Cite `Logan E Adams (2026). AI citation -> funnel/revenue. Clear Cited. https://clearcited.com/research/s-citation-funnel/`
Own-brand Reddit test (disclosed participation only)	Tier 2	proposed — blocked: Reddit stays MANUAL by policy — participate by hand (do NOT arm automation)	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	growth_retainer, scale_retainer…	Cite `Logan E Adams (2026). Own-brand Reddit test (disclosed participation only). Clear Cited. https://clearcited.com/research/s-reddit-transparent/`
Directory causal test across clients	Tier 2	proposed — blocked: Reach 3 clients with sealed T0 baselines — 3 studies cannot commence below it	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	addon_directory_boost, addon_directory_growth…	Cite `Logan E Adams (2026). Directory causal test across clients. Clear Cited. https://clearcited.com/research/s-directory-client-causal/`
PR placement -> citation lift	Tier 1	proposed	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	addon_digital_pr, scale_retainer…	Cite `Logan E Adams (2026). PR placement -> citation lift. Clear Cited. https://clearcited.com/research/s-pr-citation/`
Brand mentions vs backlinks, cleanly separated	Tier 1	proposed — blocked: Start backlink / referring-domain monitoring (discovery lag compounds)	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	addon_backlink_domain, addon_digital_pr	Cite `Logan E Adams (2026). Brand mentions vs backlinks, cleanly separated. Clear Cited. https://clearcited.com/research/s-mentions-vs-backlinks/`
Multi-baseline SCED across >=3 clients	Tier 2	proposed — blocked: Reach 3 clients with sealed T0 baselines — 3 studies cannot commence below it	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	growth_retainer, scale_retainer…	Cite `Logan E Adams (2026). Multi-baseline SCED across >=3 clients. Clear Cited. https://clearcited.com/research/s-multibaseline-sced/`
State of AI Visibility (the annual flagship)	Tier 3	proposed	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	measured, growth_retainer…	Cite `Logan E Adams (2026). State of AI Visibility (the annual flagship). Clear Cited. https://clearcited.com/research/s-state-of-ai-visibility/`
Citation-source composition BY VERTICAL	Tier 3	proposed	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	measured, full_audit	Cite `Logan E Adams (2026). Citation-source composition BY VERTICAL. Clear Cited. https://clearcited.com/research/s-source-composition-vertical/`
The leaderboard halo effect	Tier 3	proposed	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	house authority	Cite `Logan E Adams (2026). The leaderboard halo effect. Clear Cited. https://clearcited.com/research/s-leaderboard-halo/`
Ghost citations on Perplexity and Claude — a replication and extension	Tier 1	proposed — blocked: Repair citation extraction on 2 of the 5 engines before any citation study runs	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	house authority	Cite `Logan E Adams (2026). Ghost citations on Perplexity and Claude — a replication and extension. Clear Cited. https://clearcited.com/research/s-ghost-citation-replication/`
Classic rank vs AI citation, on the same questions	Tier 1	proposed — blocked: Commit the first Signal Ledger cycle — 11 registered studies read it; Repair citation extraction on 2 of the 5 engines before any citation study runs	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	measured, full_audit…	Cite `Logan E Adams (2026). Classic rank vs AI citation, on the same questions. Clear Cited. https://clearcited.com/research/s-rank-vs-citation/`
Developer adoption vs AI recommendation	Tier 1	proposed — blocked: Commit the first Signal Ledger cycle — 11 registered studies read it	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	measured, growth_retainer…	Cite `Logan E Adams (2026). Developer adoption vs AI recommendation. Clear Cited. https://clearcited.com/research/s-devtool-signals/`
Directory presence vs AI visibility, across a cohort	Tier 1	proposed — blocked: Commit the first Signal Ledger cycle — 11 registered studies read it	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	addon_directory_boost, addon_directory_growth…	Cite `Logan E Adams (2026). Directory presence vs AI visibility, across a cohort. Clear Cited. https://clearcited.com/research/s-directory-presence-xsec/`
Off-site brand signals vs AI visibility, replicated openly	Tier 1	proposed — blocked: Commit the first Signal Ledger cycle — 11 registered studies read it	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	addon_digital_pr, authority_retainer	Cite `Logan E Adams (2026). Off-site brand signals vs AI visibility, replicated openly. Clear Cited. https://clearcited.com/research/s-brand-mentions-xsec/`
How Claude selects and cites	Tier 1	proposed — blocked: Commit the first Signal Ledger cycle — 11 registered studies read it; Repair citation extraction on 2 of the 5 engines before any citation study runs	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	measured	Cite `Logan E Adams (2026). How Claude selects and cites. Clear Cited. https://clearcited.com/research/s-claude-selection/`
Absorption: does a cited page's wording shape the answer?	Tier 1	proposed — blocked: Commit the first Signal Ledger cycle — 11 registered studies read it; Repair citation extraction on 2 of the 5 engines before any citation study runs	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	measured	Cite `Logan E Adams (2026). Absorption: does a cited page's wording shape the answer?. Clear Cited. https://clearcited.com/research/s-absorption/`
Does anyone actually fetch our llms.txt?	Tier 1	proposed — blocked: Commit the first Signal Ledger cycle — 11 registered studies read it	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	house authority	Cite `Logan E Adams (2026). Does anyone actually fetch our llms.txt?. Clear Cited. https://clearcited.com/research/s-llms-txt-fetch/`
Three vendors, three numbers, one query set	Tier 1	proposed — blocked: Commit the first Signal Ledger cycle — 11 registered studies read it	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	house authority	Cite `Logan E Adams (2026). Three vendors, three numbers, one query set. Clear Cited. https://clearcited.com/research/s-method-reconciliation/`
What an AI Overview costs us in clicks	Tier 1	proposed — blocked: Commit the first Signal Ledger cycle — 11 registered studies read it	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	measured, addon_technical_seo	Cite `Logan E Adams (2026). What an AI Overview costs us in clicks. Clear Cited. https://clearcited.com/research/s-aio-ctr-house/`
What our own cadence changes did	Tier 1	proposed — blocked: Commit the first Signal Ledger cycle — 11 registered studies read it	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	house authority	Cite `Logan E Adams (2026). What our own cadence changes did. Clear Cited. https://clearcited.com/research/s-cadence-effects/`
Which of our formats get engaged with — and which get cited	Tier 1	proposed — blocked: Commit the first Signal Ledger cycle — 11 registered studies read it	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	measured, addon_blog_post	Cite `Logan E Adams (2026). Which of our formats get engaged with — and which get cited. Clear Cited. https://clearcited.com/research/s-format-effects/`
Does WHO ASKS change the answer? (persona phrasing)	Tier 1	proposed — blocked: Repair citation extraction on 2 of the 5 engines before any citation study runs	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	measured, full_audit…	Cite `Logan E Adams (2026). Does WHO ASKS change the answer? (persona phrasing). Clear Cited. https://clearcited.com/research/s-persona-phrasing/`
Does WHERE YOU ASK FROM change the answer? (US vs CA vs INTL)	Tier 1	proposed — blocked: Repair citation extraction on 2 of the 5 engines before any citation study runs	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	measured, growth_retainer…	Cite `Logan E Adams (2026). Does WHERE YOU ASK FROM change the answer? (US vs CA vs INTL). Clear Cited. https://clearcited.com/research/s-region-variance/`
Does READING LEVEL change whether a page is cited?	Tier 1	proposed — blocked: Commit the first Signal Ledger cycle — 11 registered studies read it	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	measured, full_audit…	Cite `Logan E Adams (2026). Does READING LEVEL change whether a page is cited?. Clear Cited. https://clearcited.com/research/s-reading-level-citation/`
Does the VOICE change engagement, holding the finding constant?	Tier 1	proposed — blocked: Commit the first Signal Ledger cycle — 11 registered studies read it	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	measured, growth_retainer…	Cite `Logan E Adams (2026). Does the VOICE change engagement, holding the finding constant?. Clear Cited. https://clearcited.com/research/s-voice-engagement/`
Does the SEND SLOT change opens and clicks, by region and segment?	Tier 1	proposed	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	measured, growth_retainer…	Cite `Logan E Adams (2026). Does the SEND SLOT change opens and clicks, by region and segment?. Clear Cited. https://clearcited.com/research/s-edition-timing/`
Earned vs owned: which of OUR surfaces do engines cite?	Tier 1	proposed — blocked: Repair citation extraction on 2 of the 5 engines before any citation study runs	not yet pre-registered — this study has made no commitment to check	no study-level DOI yet — this readout has not been deposited. The Index dataset and the measurement protocol each carry one; a study's citation never borrows them.	measured, full_audit…	Cite `Logan E Adams (2026). Earned vs owned: which of OUR surfaces do engines cite?. Clear Cited. https://clearcited.com/research/s-earned-vs-owned-citation/`

Each study names the SKU it is the evidence base for — see the Evidence section on the relevant pricing page — or is marked house authority (a brand/category asset, not sales copy).

Want your own category measured?

A free teardown runs your buyer prompts across every engine.

Get a free teardown

Open data & pre-registration

Clear Cited is independent AEO/GEO measurement — a research practice that sells audits. Every figure we publish carries its run count and its interval, and the data behind it is deposited, versioned and citable.

Dataset DOI: DOI 10.5281/zenodo.21612952 (opens in a new tab)
This version: 2026.07.1 - DOI 10.5281/zenodo.21614037 (opens in a new tab)
Measurement protocol: DOI 10.5281/zenodo.21614889 (opens in a new tab)
Kaggle: the dataset on Kaggle (opens in a new tab)
Hugging Face: the dataset on Hugging Face (opens in a new tab)
Pre-registration: the OSF project (opens in a new tab)

How we measure