70% of customers walk away from hotel AI that forgets them

A Twilio research piece for the APAC region, picked up by Hospitality Net, points out that 70% of consumers have abandoned an AI interaction because the system lost context. Translated to the hotel: a guest starts a chat, switches channel, comes back, and the bot greets them like a stranger. That is where the conversation dies and, often, the booking with it.

The study's diagnosis is spot on. The root is not the language model, it is guest data scattered across the PMS, CRM, booking engine and the sales team's inbox. Nobody stitches that together in real time, so the AI works blind. The fix they propose, a governed Guest Memory Layer, feels right to me: fewer new tools and more tidying up what already exists.

My read is blunt. Selling AI without first solving a unified guest identity is like putting wheels on a car with no engine. The opening is for small hotels, which can build that layer without leaning on a big integrator and win margin with fewer repeated questions and more direct conversion.