The Index
The Business Aviation AI Visibility Index
A monthly, logged, reproducible measurement of which business aviation companies are named when buyers ask AI — 120 public prompts, four engines through their official APIs, five repeats, 2,400 runs a month — published quarterly with the derived run data. It measures the models' retrieval-augmented APIs, not the consumer products, and says so on every table. Edition 1 publishes in January 2027 on October–December 2026 runs.
- Edition
- 0 (pre-launch)
- Panel
- 120 prompts across 7 segments
- Engines ranked
- OpenAI API + web search · Gemini API + Search grounding · Perplexity Sonar API · Claude API + web searchAPI-measured, not the consumer products · 2,400 logged runs a month
- Repeats
- 5 per prompt per engine per month
- Runs per month
- 2,400 logged API runs (120 × 4 × 5)
- Calibration sample (not ranked)
- 20 prompts × 2 runs, hand-run each quarter across AI Overviews, Copilot and the ChatGPT app; reported as an agreement rate, not ranked
- Monthly runs required before edition 1
- 3
- Runs logged
- 0
- First run
- October 2026
- Edition 1
- January 2027
- Cadence
- Monthly runs · quarterly public editions
No rankings are published until edition 1. We do not publish single-run results.
Measured through official APIs OpenAI API + web search · Gemini API + Search grounding · Perplexity Sonar API · Claude API + web search — the models' retrieval-augmented APIs, not the consumer products. How each surface is measured
Why an index
Every existing measure of AI visibility in private aviation is either a single-day snapshot of 14 prompts, 1 a modeled score whose methodology states "no logged query runs", 2 or a single agency's 25-brand audit. 3 None covers FBOs, MROs, management, sales, fuel or vendors. None publishes its prompt list. None reports run-to-run variance, which independent research shows is severe. 4 This Index does all of that, and publishes the run data where provider terms permit.
What is measured
Each term links to its glossary entry.
- Named
The company is mentioned by name in the answer text.
- Cited
A page on the company's domain is listed as a source.
- Share of voice
Named runs ÷ total runs for a prompt cluster and engine, expressed as a percentage with a 95% interval.
- Citation share
The share of cited domains in a prompt cluster that belong to the company.
- Position credit
1.00 first named, 0.85 second, 0.60 third, 0.35 fourth or later, 0 not named — reported alongside, never instead of, share of voice.
- Accuracy flag
A logged instance of an engine misstating the company's fleet, rating, base or certificate.
- Variance
The spread of "named" across the five repeats of the same prompt; published per prompt.
- Calibration sample
A quarterly, hand-run set of 20 panel prompts × 2 runs entered by an analyst in a clean US browser profile in the ChatGPT app, Google AI Overviews/AI Mode and Microsoft Copilot, logged with screenshots; reported only as the rate at which the app named the same companies as the API engine, never as a ranking.
The panel
The panel is public. It is built from the Prompt Library through the Buyer Question Program (consented interviews, the public opt-in form, secondary sources), balanced across seven segments and four buying stages, tagged by region and mission, and frozen for each quarter; changes are proposed monthly, applied only at quarter boundaries, and logged with reasons. Clients may not propose prompts. Edition 1 is US-only; UK/EU and Middle East cuts arrive with Edition 3 (July 2027). The vendors and software segment excludes the charter sourcing, booking and payments category in Edition 1 — see the conflict policy. Editorial independence and conflicts of interest
Panel composition · 120 prompts · segment × buying stage · Edition 1
| Segment | Discover | Compare | Verify | Decide | Total |
|---|---|---|---|---|---|
| Charter operators | 21 | 2 | 15 | 2 | 40 |
| Charter brokers | 5 | 1 | 6 | 3 | 15 |
| FBOs | 1 | 5 | 12 | 2 | 20 |
| MROs | 7 | 2 | 5 | 1 | 15 |
| Management & sales | 2 | 1 | 8 | 4 | 15 |
| Fuel | 0 | 2 | 3 | 0 | 5 |
| Vendors & software | 4 | 3 | 3 | 0 | 10 |
| Total | 40 | 16 | 52 | 12 | 120 |
Panel composition · by region tag
| US-East | US-West | UK-EU | Middle East | Any |
|---|---|---|---|---|
| 19 | 22 | 9 | 2 | 68 |
Edition 1 is US-only; region is fixed per run.
Engines and measurement surfaces
API-measured: the models' retrieval-augmented APIs, not the consumer products
Engines and measurement surfaces · the four API engines are the only ones ranked
| Engine | How it is measured | Ranked? | Label used in tables |
|---|---|---|---|
| Ranked · official, paid APIs | |||
| OpenAI (GPT family)openai api + web search, ranked api engine | OpenAI Responses API with the web_search tool, default settings; user_location US unless a region cut; fresh call per run | Yes | OpenAI API + web search |
| Google Geminigemini api + search grounding, ranked api engine | Gemini API with Google Search grounding; fresh call per run | Yes(verbatim text published only if Google's grounding terms permit storage and display; otherwise derived data only) | Gemini API + Search grounding |
| Perplexityperplexity sonar api, ranked api engine | Sonar API; fresh call per run | Yes | Perplexity Sonar API |
| Anthropic Claudeclaude api + web search, ranked api engine | Claude API with the web search tool; fresh call per run | Yes | Claude API + web search |
| Not ranked · hand-run calibration sample | |||
| Google AI Overviews / AI Modeai overviews (manual sample), hand-run sample, not ranked | no official API; quarterly hand-run calibration sample, 20 prompts × 2 runs entered by an analyst in a clean US browser profile, logged with screenshots | No — calibration sample, agreement rate only | AI Overviews (manual sample, n=40) |
| Microsoft Copilotcopilot (manual sample), hand-run sample, not ranked | no official API; same calibration sample. Candidate for Edition 2: "Azure OpenAI + Grounding with Bing Search" as an additional API-measured engine, labeled as a Bing-grounded proxy, not as Copilot | No — calibration sample, agreement rate only | Copilot (manual sample, n=40) |
| ChatGPT consumer appchatgpt app (manual sample), hand-run sample, not ranked | included in the calibration sample so readers can see how far API results diverge from the app | No — calibration sample, agreement rate only | ChatGPT app (manual sample, n=40) |
Measured through official APIs OpenAI API + web search · Gemini API + Search grounding · Perplexity Sonar API · Claude API + web search — the models' retrieval-augmented APIs, not the consumer products. How each surface is measured
We rank only what we can measure through official, paid APIs, and we label it as the models' retrieval-augmented APIs, not the consumer products. The API answers differ from the apps — different system prompts, model routing, no personalization, no logged-in state — which is a limitation we publish and also why the Index is reproducible: anyone with API keys can re-run the panel. We do not scrape consumer apps, and we do not launder scraping through a vendor for a published benchmark. A Bing-grounded API engine is a candidate for Edition 2, labeled as a Bing-grounded proxy, not as Copilot.
Schedule
- RunsSpread across each month, five per prompt per engine; the calibration sample is hand-run in the last month of each quarter.
- QA and attribution reviewFirst week of the following month.
- Subscribers' previewSeven days before publication.
- Public editionsJanuary, April, July, October.
- Edition 1January 2027, on October–December 2026 runs.
How to be included
Every company named in any run is in the Index automatically; there is no fee to appear, to be positioned, or to change the panel. The Index Benchmark subscription adds a private monthly cut — it does not change how anyone is measured. The Index Benchmark subscription
To dispute an attribution or correct a factual error in a published edition, use the contact form. Contact form: correction to the Index Disputes and outcomes are logged publicly in the corrections log.
Corrections log
No corrections logged. Disputes and outcomes for each edition will appear here.
Independent review
Each public edition is reviewed by an independent reviewer — a working journalist or academic with no commercial relationship to us, our clients, aisba.ai or Tuvoli — who checks adherence to this methodology, recomputes a sample of scores from the private raw log, and signs a review note published with the edition. The reviewer is paid a fixed honorarium per edition and nothing else.
Data publication
Each edition publishes a web page per segment, a PDF, and a CSV dataset containing, for every run: prompt, engine, model, API parameters, timestamp, region, extracted entities, positions and cited domains — the derived data anyone needs to recompute the scores. Verbatim answer text is included only for engines whose API terms permit publication (confirmed per engine before this methodology went live: derived data only for all engines until confirmed); otherwise it is retained privately and available to the independent reviewer under NDA. Names of individuals in answers are redacted before publication; prompts never name individuals; raw answers are retained for 24 months and then deleted. The license is stated on each edition; our intention is CC BY 4.0 for the derived data and aggregate tables.
Cite as: "AI Says You, Business Aviation AI Visibility Index, Edition n, {month year}, aisaysyou.com/ai-visibility-index".
Journalists may republish aggregate tables with attribution; we will supply CSV on request.
Limitations
- Prompt wording is ours; a differently worded panel would produce different lists.
- The Index measures retrieval-augmented APIs, not the consumer apps: different system prompts, model routing, no personalization, no logged-in state. The calibration sample shows how far they diverge; it does not close the gap.
- Google AI Overviews, AI Mode, Copilot and the ChatGPT app are not ranked, because they cannot be measured reproducibly without scraping.
- Enterprise Copilot and ChatGPT Enterprise instances, where much B2B research happens, cannot be observed at all.
- Edition 1 is US-only; region is fixed per run.
- Five repeats distinguish patterns from noise; they do not remove it.
- Being named is not being chosen.
Charter
No pay for inclusion, position or panel changes. Markers on every table for current and past-12-month clients, Index Benchmark subscribers and design partners who received free or discounted work. The founder's position regarding Tuvoli is stated, and the charter sourcing, booking and payments category is excluded from the vendors segment while that position is unconfirmed or an interest exists. Derived run data with every edition; verbatim text where terms permit. Independent review with recomputation. Public corrections log. Hosted only on aisaysyou.com, never on aisba.ai.
Editorial independence and conflicts of interest
aisaysyou.com is operated separately from aisba.ai, the members-only community created at the AI Symposium for Business Aviation (AISBA) 2026, organized by Tuvoli, NATA and Embry-Riddle Aeronautical University. 56
Tuvoli, a charter platform, is a co-founder of AISBA. 5 Its business is charter sourcing, booking and payments.
The founder's position regarding Tuvoli will be stated here once confirmed in writing to aisba's co-founding partners. Until then, every restriction that would apply if the founder had an interest applies: no charter sourcing, booking or payments platform is accepted as a client, and that category is excluded from the Index.
The founder's other commercial roles are listed here: to be confirmed; until then the rule below applies to any company in which the founder holds a role or shareholding. Any company in that list, and any company that competes directly with it, is (a) excluded from the public Index segment it trades in, with the exclusion stated in every edition; (b) not accepted as a client; (c) not the subject of any comparison page we publish. Where an exclusion would leave a segment table incomplete, the independent reviewer scores that segment blind and says so in the review note.
Want your own monthly cut from run one?
The Index Benchmark subscription: your brand and up to ten competitors, prompt by prompt, engine by engine, from the first run.