Brand share in AI travel answers depends on your prompt list

Only 9% of the Google AI Overviews prompts in BrightEdge’s holiday travel panel carry booking intent, and inspiration and comparison questions make up most of the rest. The finding comes from travel research by BrightEdge, which sells enterprise SEO and AI search visibility tracking, and was covered by Search Engine Journal on October 31.

That intent mix shapes every brand-share number in the same study. Booking platforms appear in 38% of the booking-intent answers BrightEdge tracked and 12% of the inspiration ones, so a single brand on a single engine can post two very different share figures depending on how the prompt list was built. Weight a prompt set toward transactional phrasing and Expedia’s share climbs without anything changing in how Google assembles answers.

Prompt wording moves the same figure further than intent does, in BrightEdge’s tracked Google AI Overviews answers:

Tracked prompts	Booking platform appears
Booking intent (9% of the panel)	38%
Inspiration	12%
Budget wording (“cheap”, “deal”, “budget”, “last-minute”)	78%
Answers without that wording	17%

BrightEdge attaches a caveat to the budget-wording row, noting that the pattern is an observed association and not evidence that using those words causes a brand to appear. The caveat is worth keeping, because the obvious misreading is to write “cheap” and “last-minute” into page copy and expect a lift. Expedia turns up in 44% of those budget-related answers, which is a fact about which prompts favour large booking platforms rather than a wording trick anyone can copy.

The category leader figures carry a second condition. BrightEdge excludes non-travel platforms such as YouTube and Reddit from its category rankings, so Hilton leading AI Overviews for hotels at 28.3% is a share among tracked hotel brands, not a share of the answer text a traveller reads. A brand can hold a quarter of that ranking while the answers it appears in lean on a Reddit thread or a YouTube video.

Leaders also split by engine, with Hilton ahead on AI Overviews and Marriott ahead on ChatGPT in the same hotel category. No sample size appears in the passages Search Engine Journal quotes, and BrightEdge calls the figures share of voice rather than rankings, since AI answers have no ranking positions to count. Given that, a few points between two brands is not a result, and a figure from one vendor’s panel will not reconcile with another’s, the same way AI Overview coverage measurements disagree by 20 points depending on the tracker.

The opening for smaller sites sits in the prompts the big brands do not win. YouTube, Reddit and AAA, the American motoring club, appear more prominently in inspiration questions, and BrightEdge did not offer a reason why. Roger Montti, who wrote Search Engine Journal’s piece, argues travel bloggers should take that ground on expertise, which has never been Reddit’s strength. The narrower conclusion the data supports is that inspiration prompts are where a small travel site can appear at all, and a prompt set built from booking phrasing will never show whether it does.

What to do

• Bucket your prompt set by intent and report share within each bucket. One blended share-of-voice number across a mixed panel tracks your prompt list as much as your visibility.
• Ask your vendor what sits in the denominator. If platforms are excluded from category rankings, as they are here, a category-leading percentage is a share among competitors you chose.
• Track budget-wording prompts separately rather than folding them in. They behave differently enough that mixing them hides both patterns.
• Keep engines apart and do not average them. The hotel leader is not the same brand on AI Overviews and ChatGPT.
• Check what counts as an appearance in your tool, since a brand can be linked in an answer without being named in the text.