HN Original: Booked in Seconds, Inside ChatGPT: A Conversation with Apaleo at ITB Berlin

At ITB Berlin, Simone Puorto spoke with Uli Pillau, Founder of Apaleo, about open platforms, what agentic AI actually means in practice, and why the hotels moving fastest on this are not the large chains. The conversation ended with a live demonstration of a hotel room being booked entirely inside ChatGPT. It took a few moments.

Open from the start

Apaleo was built on a deliberate decision not to be a closed PMS. Every capability on the platform is available via API to the outside world. Partners and developers can connect, build and deploy on top of it without restriction. Apaleo does not develop its own apps. It provides the infrastructure and lets an ecosystem of partners build on it.

That openness is what makes the agent conversation real rather than theoretical. When 100 percent of a platform’s functionality is exposed, agents can act on it. When it is not, they cannot.

What an agent actually is

Uli walked through the concept simply. An agent is anything that takes over work a human currently does manually, freeing hotel staff to focus on guests rather than procedures. There are two broad categories: guest-facing agents, which interact directly with travellers, and staff-facing agents, which automate internal workflows.

A staff-facing agent handles something like extending a reservation. Instead of a staff member navigating through screens and updating fields manually, they type the instruction and the agent executes it in seconds. A guest-facing booking agent goes further: a guest has a conversation inside ChatGPT, availability and pricing are pulled from Apaleo in real time, the guest selects a room, enters their details by typing or by voice, and a payment link is generated. The reservation lands directly in Apaleo. The guest never leaves the chat.

This is not a roadmap item. Apaleo currently has 15 to 20 hotel groups deploying agents in production. The observation Uli made about who is moving fastest is worth noting: it is not the large chains, which move slowly by structure. It is the smaller, more agile operators who are testing, iterating and making it work.

The live demo

Partway through the conversation, Uli demonstrated the ChatGPT booking flow live. A guest query entered the chat. Apaleo responded with real-time availability and pricing. The guest selected a room, confirmed details, received a secure payment link and completed the booking. The reservation appeared in Apaleo. The entire sequence took a few moments, and none of it was staged.

The significance is not just the speed. It is that the booking happened entirely inside a conversational interface, with no separate booking engine, no redirect to a hotel website, no form to fill in. The guest experience was continuous from search to confirmation.

Human and digital labour, side by side

Uli is clear that agents are not replacing people in hospitality. The industry is built on human interaction and that does not change. What agents replace is the procedural, repetitive back-office work that currently occupies staff time and gets in the way of that interaction. His suggestion for any hotel evaluating this: start by asking whether the infrastructure is in place to support agents, then go through every procedure in the operation and ask whether an agent could take it over. In most cases, the answer today is yes, and the speed of development means the answer will only become more comprehensive.

The revolution Uli sees coming is, by his own description, bigger than anything he has witnessed in hotel technology. The limiting factor is not the agents. It is whether the underlying platform is open enough to support them.