answer engine optimization and AI search visibility — dataset

About this freeze

Frozen measurement export for answer engine optimization and AI search visibility (measured 27 July 2026). Primary evidence — re-derivable from the published methodology. Negatives (cited: false) stay in the freeze and count in rates. The HTML table is a short preview; the full freeze is in the CSV/JSON downloads.

Each row is one direct answer-engine check. We asked the category question verbatim through the engine's live API and recorded whether the answer cited any sources — cited: true means at least one source URL; cited: false is a measured negative, not missing data. Measurement date: 27 July 2026.

Engines and sampling rules (including any variance / test-retest panel) are disclosed on the linked methodology page for “answer engine optimization and AI search visibility.” This dataset page freezes the rows; it does not recompute the study narrative.

Full methodology

Findings from this freeze

Declarative facts computed from this freeze only — not a substitute for the study article.

Overall freeze

• In this freeze, 75 of 92 checks (81.5%) recorded at least one cited source.
• 17 of 92 frozen checks recorded no citations (negatives kept).
• This freeze covers 24 distinct queries.

Citation rates by engine

• chatgpt: 7 of 24 runs (29.2%) recorded ≥1 cited source.
• gemini: 24 of 24 runs (100%) recorded ≥1 cited source.
• perplexity: 44 of 44 runs (100%) recorded ≥1 cited source.

Summary statistics

Computed from the frozen rows and rollup sample counts. No figures are invented at render time.

• Rows92
• Distinct queries24
• Enginesperplexity, gemini, chatgpt
• Checks with ≥1 citation75 of 92 (81.5%)
• Checks with no citations17
• Observation dates2026-07-27
• Queries in design24
• Checks recorded92
• Checks planned92
• Variance panel groups10

Query coverage

Distinct category questions present in this freeze.

Questions in this freeze

• what is answer engine optimization
• AEO vs SEO what is the difference
• what is generative engine optimization
• what are the best AEO tools
• best tools to track ChatGPT citations
• platforms that measure share of voice in ChatGPT
• AEO platforms for marketing agencies
• how do brands get cited in ChatGPT
• how to appear in Google AI Overviews
• how to measure AI search citations
• how to audit a website for AI readiness
• how to build entity authority for AI search
• how to track competitors in AI search engines
• how agencies deliver AEO services to clients
• how Reddit and review sites affect AI citations
• how schema and entities affect AI citations
• why brands are not cited by AI search engines
• why ChatGPT recommends some brands over others
• common mistakes in AI search optimization
• what content formats get cited by AI answer engines
• what sources do AI answer engines trust most
• what makes a page extractable for AI answers
• how to prove AI citation improvement to a client
• metrics that matter for AI visibility reporting

Column dictionary

Field-by-field map of the freeze. The cited column is the direct citation detector: true means at least one source URL; false means none were cited.

Core detectors

cited: true / false

Whether this check recorded at least one cited source URL (true) or none (false). False rows are negatives and stay in the freeze.

variance panel

A fixed subset of questions re-asked the same day (k repeats) so citation flip rates are measured, not guessed.

check_group_id

Groups k-repeat draws of the same query on the same engine for test-retest comparison.

run_seq

Repeat-draw index within a same-day variance panel (1 = first draw).

All columns

cited

Whether this check recorded at least one cited source URL (true) or none (false) — negatives are data.

query

The category question put to the answer engine.

engine

Which answer engine produced this observation (e.g. perplexity, chatgpt, gemini).

run_at

ISO timestamp when this engine check was recorded.

run_seq

Repeat-draw index within a same-day variance panel (1 = first draw).

source_urls

Pipe-separated full source URLs cited in the answer, when any.

check_group_id

Groups k-repeat draws of the same query on the same engine for test-retest.

source_domains

Pipe-separated domains cited in the answer, when any.

How to use this data

1. Download the CSV or JSON for independent re-analysis — every row is frozen at measurement time.
2. Read the methodology page for sampling rules, engines, variance panels, and limitations.
3. Cite the study article for the narrative findings; cite this dataset when quoting row-level figures.
4. Treat cited: false (or empty source columns) as measured negatives, not missing data.

Measurement preview

Showing 8 of 92 rows with the scannable columns only (source URL lists stay in the download). Download CSV or JSON for the complete freeze.

Full freeze: CSV · JSON