Travellers ask AI where to book. Does it send them to you?
The trip used to start on a search results page. Now it starts with an assistant naming two or three sites - and the booking goes to one of them.
We asked both assistants what a traveller would ask.
ChatGPT
Asked: Which website is best for booking cheap flights?
Google Flights is the best place to start for most people. It makes comparing dates, nearby airports, and price trends easy, and it offers price alerts.
For the lowest possible fare, I’d also check Skyscanner and KAYAK—recent comparisons found that the cheapest result varies by route.
Best strategy: Google Flights → Skyscanner → KAYAK → check the airline’s own website before booking.
3 brands named
Google Flights
Skyscanner
KAYAK
Claude
Asked: Which website is best for booking cheap flights?
There's no single "best" site—it really depends on what you're looking for, but a few consistently rise to the top for finding cheap flights:
Top picks:
Google Flights – you can find cheap flight deals from United States to anywhere in the world on Google Flights by entering your departure city, choosing Anywhere as the destination, and selecting Explore. It also has excellent price tracking tools that show you the best time to book to get the cheapest price, and can notify you when prices drop.
6 brands named
Google Flights
Skyscanner
Kayak
Momondo
Priceline
Hopper
Real answers - asked from the United States on 25 September 2026. Shortened; the brand count covers the whole answer.
The gap
ChatGPT named 3 brands. Claude named 6. Is yours one of them?
Someone who asks this never opens a results page. They get a short list, pick from it, and the brands that were not on it never had a chance.
01
The list is short
An answer names a handful of brands, not ten blue links. Being the eleventh result used to be visible. Being the eleventh brand is invisible.
02
Each assistant reads differently
The two answers above come from the same question and name different brands, because the assistants read different sources. Winning one does not win the other.
03
It changes every week
Answers move as models update and new pages get cited. A single check tells you where you stood once; it does not tell you when you dropped out.
Check your brand
Above is one question. Check all of them.
One report answers it today. Daily monitoring shows whether it stays that way - and what to do when it does not.
AI Search Insight
One-time report
Find out once where you stand in AI answers.
AI recommends companies in your industry. The report checks whether yours is among them - 30 real customer questions, asked in five assistants. No account and no subscription: you pay for the report.
An AI visibility platform for marketing teams and agencies that want the work done, not another dashboard. Five assistants measured every day, and an agent that acts on what they say.
You are competing with aggregators for your own inventory.
Demand arrives months before the purchase and disappears for half the year, so content has to be planned against a calendar rather than a backlog.
Travel demand starts at inspiration - where, when, is it worth it - long before anyone searches for a specific property or route. That stage is where models are now heavily used, and where aggregators have historically owned the search results.
The organic opportunity is in the specifics only an operator knows: real conditions, real timing, real constraints. That is also exactly the material models cite, because aggregators cannot generate it.
What it covers
Seasonal content planning tied to the booking curve
Inspiration-stage content that aggregators cannot write
Destination and route architecture at scale
Multi-market and multi-language structure
AI visibility on planning and recommendation prompts
How it works
01
Build the calendar backwards from the booking curve
People research months before they book and stop entirely for half the year. Content that goes live when the season opens has already missed it, so the plan runs against the curve rather than against the quarter.
02
Own the inspiration stage, where aggregators are weak
Where, when, is it worth it, what is it actually like in March. Aggregators cannot generate that - they have inventory, not experience - and it is the stage where models are now heavily used.
03
Write the specifics only an operator knows
Real conditions, real timing, real constraints, including the ones that cost you a booking. That is the material that gets cited, precisely because nobody selling inventory would publish it.
04
Structure destinations before scaling them
Destination and route pages multiply fast and become thin faster. The architecture and the quality floor get decided before the first hundred exist, not after a thousand do.
What you get
Content calendar aligned to the booking curve, not the fiscal one
Inspiration-stage content aggregators cannot produce
Destination and route architecture with a quality floor
Multi-market and multi-language structure
AI visibility on planning and recommendation prompts