Where is the real limit of AI in hotels? The CEO of Neobookings explained it at Global Revenue Forum Madrid 2026

José María answered with an example any hotelier will recognise. A thirty-room hotel has a chatbot that answers questions, an assistant that drafts the replies to reviews and a switchboard that handles calls. Three tools, three providers, three different knowledge bases. The day a rate, a cancellation policy or the opening hours of a service change, someone has to remember to update all three. But hardly anyone ever remembers all three. That is when the problems start.

But it is a much simpler problem than it looks: the information lives scattered across applications that do not talk to each other.

Two levels of adoption, and only one is solved

Technically, José María explained, almost anything a hotel can think of can already be automated; today the power that technology offers is practically unlimited. What is missing is for all that capability to speak the same language inside the operation.

Most hoteliers already use ChatGPT or Claude for one-off tasks such as writing an email, generating an image, putting a campaign together. That adoption has been fast and is practically universal across the industry. In that sense, there is nothing new.

But there is a second level that almost nobody has solved yet. It is no longer enough to ask AI something and get an answer back. Artificial intelligence needs access to all the information in the operation in order to decide and act on its own. That is where real coordination between channels comes in: website, phone call, payment, team.

Other panellists agreed from different angles. Tomás Rodicio, from Nethits Telecom, pointed out that the hotel industry has been working with APIs and connected systems for years, so part of the data is already organised; the challenge lies in automating routine processes that used to be done by hand. Judit Segura, from Spirit Hotels, described the opposite case: they automated online check-in and messaging across several channels without first defining the communication flow, and the result was chaotic for the guest rather than an improvement. They learned, she said, that strategy comes first and the tool comes second.

The channel manager analogy

José María explained it with an example the whole industry knows well, the channel manager. Nobody enters a different availability by hand into each OTA; there is a single source that distributes that information outwards.

The same applies to AI in the operation. If the chatbot, the switchboard, the booking engine and the CRM each pull data from their own base, every change has to be replicated by hand in each system. If they all connect to a single source of information, through open APIs or MCPs, that change propagates on its own.

Making this change does not require centralising everything at once. José María turned it into a simple exercise: for one week, have each department note down which task takes up the most time: reservations, revenue, front desk, marketing and so on. That inventory will reveal, without any major investment, where to automate first and which systems should start talking to each other.

Before adding one more tool, the first step is to decide where all that information is going to live from now on.

Do you want your channels to work from the same information? Let us talk about where to start.