AI Says YouAI Says You[1]
Menu

Service · 04

Index Benchmark subscription

$500 per month or $5,000 per year · no services included

Your own monthly cut of the Business Aviation AI Visibility Index: your brand and up to ten named competitors, prompt by prompt, engine by engine, with trend charts.

Request a subscription
Exhibit · Index status · edition 0
API-measured · not the consumer productsA plate of 120 prompt columns and 4 engine rows (OpenAI API + web search, Gemini API + Search grounding, Perplexity Sonar API, Claude API + web search), 5 repeats each — 2,400 logged API runs per cycle; 0 logged so far.API-measured · not the consumer productsOpenAI API + web search×5Gemini API + Search grounding×5Perplexity Sonar API×5Claude API + web search×5120 prompts →monthly · 2,400 logged api runs · 0 of 3 monthly runs loggedAPI-measured · not the consumer productsA plate of 120 prompt columns and 4 engine rows (OpenAI API + web search, Gemini API + Search grounding, Perplexity Sonar API, Claude API + web search), 5 repeats each — 2,400 logged API runs per cycle; 0 logged so far.API-measured · not the consumer productsOpenAI API + web search · ×5Gemini API + Search grounding · ×5Perplexity Sonar API · ×5Claude API + web search · ×5120 prompts →monthly2,400 logged api runs0 of 3 monthly runs loggedCalibration sample · hand-run quarterly · not rankedA separate 20 × 2 plate per surface (AI Overviews, Copilot, ChatGPT app), entered by hand each quarter and reported only as an agreement rate. Never ranked, never pooled with the API runs.Calibration samplehand-run quarterly · not rankedAI OverviewsCopilotChatGPT app20 prompts × 2 runs · n=40agreement rate only · never pooledCalibration sample · hand-run quarterly · not rankedA separate 20 × 2 plate per surface (AI Overviews, Copilot, ChatGPT app), entered by hand each quarter and reported only as an agreement rate. Never ranked, never pooled with the API runs.Calibration samplehand-run quarterly · not rankedAI OverviewsCopilotChatGPT app20 prompts × 2 runsn=40agreement rate onlynever pooled
Panel
120 prompts · 7 segments · published
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
Calibration sample (not ranked)
AI Overviews · Copilot · ChatGPT appHand-run quarterly · 20 prompts × 2 runs · agreement rate only
Monthly runs logged
0 of 3 required
Edition 1
January 2027

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

Read the methodology and the panel

Exhibit · Index status · edition 0
Who it is for
Firms that want the measurement and nothing else, including firms that will never buy services.
What you get
  • A private monthly page with your share of voice per engine and prompt cluster, prompt-level detail and the run log for runs in which you are named (where provider terms permit).
  • Up to 10 named competitors tracked the same way.
  • Trend charts from your first month.
  • Each public edition seven days before publication.
  • CSV export.
What it does not buy
Any influence over the Index. Subscribers and non-subscribers are measured identically; subscribers may not propose panel prompts; and because a subscription is a paid relationship with the Index publisher, subscribers carry the same visible marker in every Index table as clients do. The Index charter applies. The Index charter
Before edition 1
Subscribers joining before January 2027 receive monthly run data from their first run onward. Rankings are not published to anyone until three monthly runs are complete.

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

Questions about the subscription

Is our data public?

Only what each public edition publishes for every company on the panel: whether you were named, per prompt cluster and engine, as a share of runs. Your private page shows more; the public edition shows the same for all.

We will never buy services. Is this still for us?

Yes. It is designed for firms that want the measurement and nothing else.

Request a subscription