Methodology — How Cited Measures Travel AI Visibility

What we measure

The Cited Score measures how often a travel brand is mentioned, cited, and positively recommended when relevant buyers ask AI engines questions about their segment. We do not measure rank position within a single response, because position varies with every AI output. We measure visibility rate: the percentage of relevant buyer-intent prompts across a standardised battery that mention your brand.

Core principle: Position inside one AI answer is noise. Citation rate across 2,400+ standardised prompts is signal. We measure the signal.

AI engines tracked

ChatGPT

OpenAI GPT-4o and GPT-4 Turbo

Google AI

AI Overviews and Gemini

Perplexity

Perplexity Pro search

Each engine receives an identical prompt battery. Results are tracked separately and combined into the composite Cited Score. Engine-specific subscores are available in Growth and Enterprise plans so you can see where your visibility strength and gaps differ by platform.

The prompt battery

For each travel segment, Cited maintains a standardised set of buyer-intent prompts covering the questions real buyers ask when shortlisting travel brands. For Business Travel and Travel Management Companies, the base battery is 480 prompts per engine, each run 20 times. That is 28,800 data points per market per month.

General shortlisting

Best [segment] in [market]

Regional

[Segment] in [city/country]

Feature-specific

[Segment] with [capability]

Use-case

[Segment] for [company type]

Comparison

[Brand A] vs [Brand B]

Alternatives

Alternatives to [brand]

Handling non-determinism

AI engine outputs are non-deterministic: the same prompt can return different results on different runs. Each prompt in our battery is run 20 times. The citation rate for that prompt is the percentage of those 20 runs where your brand was mentioned. This volume eliminates random variation as a meaningful factor in the score.

Version tracking: We record the model version and API date for each run batch. When major model updates occur, we flag this in your score report. Model updates can shift citation patterns industry-wide, independent of anything any specific brand does.

Cited Score components

AI Mention Rate

20%

How often the brand appears in AI responses across the full prompt battery for the segment and market

Citation Rate

20%

How often the brand is cited or linked as a source, not just mentioned in passing

Recommendation Quality

20%

Whether the brand is positively recommended versus mentioned neutrally or negatively

Query Coverage

15%

The proportion of the 480-prompt battery where the brand appears at least once

Source Authority

15%

Third-party signals: reviews, press mentions, directory listings, and structured data that AI engines trust

Content Readiness

10%

How clearly and completely the website answers the buyer questions AI engines use to form answers

The Cited Score is a composite of all six components weighted as shown. Scores range from 0 to 100. A score above 70 indicates strong AI visibility in that segment and market.

Transparency and limitations

We publish our methodology openly so clients can act on it more effectively. Two limitations we are honest about:

Model updates: When AI engines update their underlying models, citation patterns can shift industry-wide independently of what any brand does. We flag these events in score reports and explain the likely impact.
Gaming risk: Once the methodology is public, brands may optimize specifically for our prompt battery. We update the battery quarterly, disclose the version with each report, and design queries to resist narrow optimization by requiring genuine breadth of AI visibility.

Get your Cited Score

We run your brand through this full methodology. Your Cited Score is delivered withwithin 24 hours of submission.

Get my score See pricing