MARKET
Spain · hotel software
Before booking a demo, a hotel manager asks AI which software to choose. Here you see where the assistants send them. If your name doesn't come up, the demo goes to someone else.
Measured on September 13, 2026 across 12 assistants: ChatGPT · Claude · Copilot · DeepSeek · Gemini · GLM · Grok · Kimi · Meta AI · Mistral · Perplexity · Qwen
10 questions
Of 27 suppliers, 5 are never named. The others show up, but not equally: a few come back in almost every answer, and those are the ones that make the buyer's shortlist. The good news is that the shortlist changes when the sites the assistants read change.
Who gets recommended, and how often.
Being named once isn't enough. Buyers ask several questions, to several assistants, and in the end they choose among the names they've heard most.
- Cloudbedsin 25 answers
- Mewsin 17 answers
- Amenitizin 14 answers
- Roomraccoonin 14 answers
- Oracle OPERAin 13 answers
- AvaiBookin 11 answers
- Little Hotelierin 11 answers
- SiteMinderin 10 answers
- Aviratoin 8 answers
- Lodgifyin 8 answers
- Smoobuin 7 answers
- Hostawayin 5 answers
- Avantioin 4 answers
- Beds24in 4 answers
- Guestyin 3 answers
- Protelin 3 answers
- Octoratein 2 answers
- Sirvoyin 2 answers
- eviivoin one answer
- LobbyPMSin one answer
- Norayin one answer
- Ulysesin one answer
The good news is that how often you show up isn't fixed: it changes with the sites the assistants read, and the measurement tells you which ones.
Every assistant has its favourites
This is where a market really makes sense: ChatGPT, Claude and Gemini don't recommend the same names. For each assistant you see the three suppliers it names most, and how often.
Claude
- 1Avirato2
- 2Cloudbeds2
- 3Mews2
Copilot
- 1Amenitiz3
- 2Cloudbeds2
- 3Little Hotelier2
DeepSeek
- 1Cloudbeds3
- 2Mews3
- 3Oracle OPERA3
Gemini
- 1Avirato2
- 2Cloudbeds2
- 3Mews2
GLM
- 1Amenitiz1
- 2AvaiBook1
- 3Lodgify1
Kimi
- 1Cloudbeds4
- 2Amenitiz2
- 3Mews2
Meta AI
- 1Little Hotelier3
- 2Oracle OPERA3
- 3Cloudbeds2
Mistral
- 1Little Hotelier2
- 2Amenitiz1
- 3AvaiBook1
ChatGPT
- 1Amenitiz2
- 2Cloudbeds2
- 3AvaiBook1
Perplexity
- 1Cloudbeds2
- 2Amenitiz1
- 3AvaiBook1
Qwen
- 1Cloudbeds2
- 2Amenitiz1
- 3Little Hotelier1
Grok
- 1Roomraccoon4
- 2AvaiBook3
- 3Cloudbeds3
You can show up often on one and not exist on another. The good news is you don't need to show up everywhere at once: the measurement tells you which assistant you're missing from, and that's where you start.
The sites the assistants read.
To choose a supplier, the assistants read industry directories, comparison sites and blogs. If your product isn't on these sites, to them it doesn't exist. Below you can see which ones each of them reads.
Claude
reads 63 sites
- booking.com
- roomraccoon.es
- amenitiz.com
- avaibook.com
- sede.agenciatributaria.gob.es
- lodgify.com
Copilot
reads 40 sites
- booking.com
- comparasoft.es
- smartorder.ai
- amenitiz.com
- hoteltechreport.com
- sede.agenciatributaria.gob.es
DeepSeek
doesn't say which sites it reads
Gemini
reads 11 sites
- booking.com
- amenitiz.com
- roomraccoon.es
- comparasoft.es
- nobeds.com
- prostay.com
GLM
reads 85 sites
- youtube.com
- roomraccoon.es
- es.hotelsync.com
- ofi.es
- prostay.com
- sentaury.com
Kimi
doesn't say which sites it reads
Meta AI
reads 13 sites
- booking.com
- siteminder.com
- baa.legal
- cuatrecasas.com
- duckduckgo.com
- hoteltechnologynews.com
Mistral
reads 22 sites
- booking.com
- amenitiz.com
- avaibook.com
- hoteltechreport.com
- sede.agenciatributaria.gob.es
- cloudbeds.com
ChatGPT
reads 24 sites
- amenitiz.com
- booking.com
- avaibook.com
- hoteltechreport.com
- sede.agenciatributaria.gob.es
- cloudbeds.com
Perplexity
reads 125 sites
- hoteltechreport.com
- amenitiz.com
- cloudbeds.com
- lodgify.com
- mews.com
- skello.es
Qwen
doesn't say which sites it reads
Grok
reads 69 sites
- booking.com
- amenitiz.com
- pyme.es
- avaibook.com
- hoteltechreport.com
- mews.com
Only one site is read by all of them: booking.com.
The good news is that these sites can be reached: a listing on a comparison site or an article on an industry blog is enough to get into an assistant's reading. The measurement tells you which ones matter for your market.
The questions people ask AI.
The ten questions buyers really ask, in their own language, from e-invoicing to support in Spanish. They stay the same at every measurement, so you can see whether you show up more or less often.
- 01¿Cuál es el mejor software de gestión hotelera en España para un hotel de 45 habitaciones?
- 02¿Qué programa sincroniza disponibilidad con Booking y actúa como channel manager sin fallos aquí en España?
- 03¿Cuál cumple al cien por cien con la factura electrónica y las normativas de Hacienda en España?
- 04¿Qué gestor ofrece soporte técnico completo en español y no en inglés o con traductores automáticos?
- 05¿Qué diferencia hay entre usar un sistema para un hostal pequeño en España y uno enfocado a un hotel grande?
- 06¿Cuánto cuesta al mes o por habitación contratar un programa de estos en España?
- 07¿Qué tipo de contrato de permanencia te obligan a firmar en España y cómo se puede cancelar?
- 08¿Conviene más usar un programa en la nube o instalado en los ordenadores del hotel en España?
- 09¿Qué otras opciones existen en España además del programa que utiliza todo el mundo?
- 10¿Cómo se puede cambiar de un programa viejo a uno nuevo en España sin perder los datos de los clientes?
If your product solves one of these problems but doesn't show up when someone asks, that question sends customers to someone else. The good news is that the measurement tells you which questions you're missing from, one by one.
Where these numbers come from.
A few rules, the same for every market. They serve one purpose: every number on this page can be checked.
- The platform writes the questions, starting from what matters to buyers in this sector. Once written they don't change, so every measurement can be compared with the one before.
- The list of suppliers is decided before measuring. If an assistant names someone who isn't on the list, the answer contains it but this page doesn't count it. If your product is missing, and we'll add it to the next measurement.
- A supplier counts as named when the assistant writes its name. A link to the site without the name isn't enough.
- Every assistant gets the same questions on the same day. An answer that doesn't arrive isn't worth zero, it counts as missing: in this measurement 4 out of 120 are missing.
- Assistants don't always answer the same way, and a single measurement can't tell a trend from a fluke. That's why the platform asks again over time for those who follow their market.
- Some assistants, like DeepSeek, Kimi, and Qwen, don't say which sites they read. That doesn't mean they read none: it means there's no way to know which.
Do you sell in this market?
Mador measures your product assistant by assistant and tells you who gets recommended instead of you. Then it works to get you there, and to keep you there once you are: it writes the texts that answer the questions you're missing from, and asks again over time, so you can see if it's working.
Leave us your product's name and a contact: the first measurement starts there.