Its average first response is approximately five seconds
_w=800_h=533.png?v=20230522122229)
Bookboost reveals 'personalised' AI Agent trained on guest profiles
Bookboost, the hospitality Customer Data Platform, announced at ITB Berlin 2026, the launch of its AI Agent.
"Unlike standalone AI chatbots that treat every guest as a stranger" the firm said Bookboost's AI Agent draws "directly" on unified guest profiles from the Bookboost CDP, so it already knows a guest's stay history, preferences, and booking context before the first message is sent.
The result is 24/7 autonomous guest service that is genuinely personalised, not just automated.
Most hotel AI tools operate in isolation. They can answer questions, but they don't know who is asking. Because
The hospitality provider has built AI Agent natively connected to the Bookboost CDP, so it has access to the same same unified guest data that powers the entire platform which it uses to generate autonomous personalised responses with the information the hotel already knows about its guests.
"The problem with most AI in hospitality isn't the AI, it's the data behind it," said Willem Rabsztyn, CEO & Co-founder.
"Any chatbot can answer 'what time is checkout.' What our AI Agent does differently is understand the full guest relationship and act on it. That's only possible because it shares the same data layer as everything else in our platform."
The firm trained the agent on five years of real hospitality guest interactions, giving it the domain knowledge hotels would otherwise spend months building.
It reads full conversational context, understands guest intent, and manages multi-step interactions from first question to resolution.
Picking up on human value still need - when a conversation requires human judgement, it transfers to the hotel team via the Bookboost Unified Inbox, with full context intact so guests never have to repeat themselves.
It's available across WhatsApp, Messenger, Instagram, email, and web chat, and supports natural multi-language responses.
Its early results have show common conversations resolved without staff involvement, with an average first response of approximately five seconds saving hours per team member per week.
