Diagnostic · Hotel assistant
An assistant / chatbot audit for your hotel
Horizonvivo reviews the conversational assistant of independent hotels in the Valencian Community and on the Costa Blanca: we check whether it really answers the guest’s questions, in their language, whether it knows the season and the area, and whether it protects personal data. An assistant that only redirects to a phone call is not helping anyone book.
Why it matters
What the sector data says
- 94M
turistas internacionales en España en 2024: una parte importante de tus huéspedes no escribe en español.
Fuente: INE, 2024
- +11%
de margen de precio asociado a subir un punto la reputación online — y la atención rápida y útil pesa en esas reseñas.
Fuente: Cornell University, Anderson 2012
- Junio
el mes en que un hotel auditado de la zona no admitía reservas por el bot pese a ser temporada alta — la fuga estaba en el asistente.
Fuente: caso de auditoría Horizonvivo, Alicante
In detail
How we review it
We put the assistant through the questions a real guest would ask: can I book here?, how much is the day pass?, until when is the pool open? We note when it answers usefully, when it replies "call reception", and when it simply does not understand. A bot that sends you to call for every question is, in practice, a contact form in disguise.
We check context-awareness. In a real case in the area, the assistant said the season ended in late August, when on that coast it runs well into September: it had been integrated without knowing it lives in Spain. We also look at language: if the guest writes in English or German, does it answer just as well?
And we review privacy: what data the assistant asks for, where it ends up and whether it exposes it. A bot that leaks personal data is not just a poor experience, it is a risk. We hand you the concrete points to fix, without replacing your tool if it works.
| Situation | Useful assistant | "Call us" assistant |
|---|---|---|
| Price question | Answers with the price | Asks you to call |
| Guest in English | Replies in their language | Answers poorly or not at all |
| Season and area | Knows the local calendar | Gives the wrong dates |
| Personal data | Asks for and stores them carefully | Risk of exposure |
FAQ
What people usually ask
Do we have to remove our current chatbot?
Not necessarily. Often the problem is the configuration or the content it was trained on, not the tool. We tell you whether it can be fixed by adjusting what is there or whether something else is worth it.
How do you check whether it leaks data?
We ask the questions that might make the assistant reveal information it should not, and observe what data it requests and how it handles it. We perform no intrusive action on your systems.
Do you test in several languages?
Yes, at least Spanish and English, and German or others if your guest mix calls for it. The guest’s language is often where the assistant falls short.
What do I get from the diagnostic?
A clear list of what answers well, what answers poorly and which concrete changes have the most impact, ordered by priority. With no unnecessary jargon.
Want to see if this is happening to you?
The first diagnostic is free and with no obligation. We review it on your hotel and tell you what we find.