Travel and OTA
Travellers still search for the trip. The answer names someone else.
BrandPulse measures both sides for an online travel brand. We read the demand in your Search Console. Then we scan what the models answer. You see the travel topics you own, and the ones an answer hands to a competitor.
An engagement starts with a paid diagnostic. There is no self-serve plan for a brand group.
The measured picture
Four numbers a travel team must hold at once
Read the order. The booked damage comes first, the honest counterweight second, and the sellable insight last.
$457M
written down on Kayak, after Google AI search changes cut organic visibility.
Source: Hospitality.today · October 2025Under 1%
of Booking Holdings room nights come from AI referrals. The share has not moved.
Source: Booking Holdings Q2 FY2026 earnings call, via Investing.com · August 2026+194%
growth in AI referrals to US travel sites. Adobe published no share of total traffic.
Source: Adobe Analytics, via Search Engine Land · May 20263.3%
citation share for Booking.com on travel questions, against 55% on hotel deal prompts.
Source: 5W Public Relations AI Visibility Index and Meltwater Online Travel Platform AI Visibility Index · March and May 2026We cite every number we publish. If a number has no source, do not trust it.
For online travel brands
What a travel group sees in the first month
The measured loss in travel is on Google, not in the chatbots. We start there.
The Google side, measured first
We read your Search Console and rank the travel queries that carry real demand. Then we show which of them no model mentions you for.
Both prompt classes
Answers name booking sites on deal prompts. The same answers name airlines, hotels and guides on trip prompts. We scan both classes.
Every brand, every market
Most groups market two or three brands and leave the rest alone. We scan each brand and each market as its own set. The quiet ones hide the blind spots.
Dutch and German prompts
Each language gets a native prompt set and a local competitor set. An English prompt set measures an English market and misses yours.
A travel competitor set
Large agencies, metasearch sites, and the airlines and hotels that sell direct. You choose the set, and we still record every other name in the answer.
A named contact
The managed programme covers the scan plan, the report and the review. One person owns it. There is no seat count and no self-serve queue.
Before you talk to us
What we will not tell you
You know this market better than any vendor does. So here is what we refuse to claim, and why.
- AI does not send you bookings today.
- Booking Holdings reports AI referrals below one percent of room nights. That share has not moved. A vendor who sells AI as a demand channel for travel is selling you a forecast.
- The +194% number has no denominator.
- Adobe measured growth in AI referrals to travel sites. Adobe published no share of total traffic. Growth of one percent is still one percent.
- We cannot make a model name your brand.
- No tool can promise that. We measure what the answers say and where the demand goes. What you publish, and how you rank, stays your work.
- We have no travel customer to show you.
- BrandPulse has no case study in travel. Nobody in this category has published a controlled lift study either. We would rather say that than show you a wall of logos.
- One answer is not a measurement.
- We ask each prompt once per scan today. A single answer is a single draw. The schedule is what makes the number mean something, so read the trend and not one week.
- The join is a match, not an identifier.
- We connect a Search Console query to a prompt by text similarity. We print the confidence on every point. You can read every match, and you can reject it.
Every number on this page carries its source and its date. If a vendor gives you a number without both, ask for both.
Questions
Questions travel teams ask
Do you run Dutch and German prompts?
Yes, in the managed programme. We build the prompt set in the language your travellers use. Each language gets its own competitor set, because the named brands change per market. An English prompt set measures an English market, and it will miss a Dutch one.
Do AI assistants send bookings today?
Almost none. Booking Holdings reported in August 2026 that AI referrals stay below one percent of room nights. We will not tell you that AI is a demand channel today. The measured loss is on Google. AI Overviews answer the question and the click never arrives. That loss is visible in your Search Console, and that is where we start.
Which competitors do you track for an online travel agency?
You choose the set. A normal set holds the large agencies, the metasearch sites, and the airlines or hotels that sell direct. We record every brand the answer names, so you also see the sites you did not think to track.
Why do the published travel indexes disagree about us?
Because visibility depends on the prompt class. On deal prompts the answers name booking sites often. On trip planning prompts the same answers name airlines, hotels and guides instead. Meltwater measured the first class in March 2026 and 5W measured the second in May 2026. We scan both classes, so you see where you win and where you disappear.
How is this different from the AI visibility tools for hotels?
Those tools measure a hotel against the agencies that resell it. Lighthouse, Cendyn and Hotelrank.ai all sell direct booking share to hotels. An agency sits on the other side of that comparison. We build the prompt set, the competitor set and the report for the agency side.
We run several brands in several markets. Can you cover them?
Yes. The managed programme covers each brand and market as its own scan, with its own prompts and its own competitors. We also read one Search Console property per brand. Most groups market two or three brands and leave the rest unattended. The unattended ones are usually where the blind spots are.
How does an engagement start?
It starts with a paid diagnostic. We scan your brands, read your Search Console, and report what the models say today. You get the prompt list, the answer text and the blind spot chart. You decide after that whether a monitoring programme is worth it.
Next step
Start with a paid diagnostic
Tell us your brands and your markets. We scan them, we read your Search Console, and we report what the models answer today. You get the prompt list, the answer text and the blind spot chart. You decide after that.
We answer by email. One person owns the programme, and there is no seat count.