Every cufflink, charger, passport and lipstick that gets left in a room ends up in a small, AI-powered detective agency. Photographed, categorised, matched against the lost-item queue — and signed for on collection. The hotel's lost property bureau, in your pocket.
The slow lane. Photographed from five angles, locked in a safe deposit, kept for the legal maximum.
The fast lane. A countdown begins on arrival. Courtesy notification before disposal — for the prescription bottle, especially.
The lane that has rules. Jurisdictional hold periods, sealed storage, disposal logged. Some things can't be returned at all.
The pragmatic lane. A short hold, then donated or discarded. The notification is courtesy, not duty.
MessageBox runs every claim and every found-item record through the same embedding model. The match isn't a keyword search — it's a vector similarity score across visual, semantic, and locational features.
That's how "leather card-holder, says JC on the front" finds the wallet that the cleaner described as "small black leather, gold lettering."
From bedside drawer to the right pocket — every state change logged, photographed, signed. Auditors love this. So does the airline that asks you to ship the wallet on.
On collection — at the concierge — the recipient signs on the device. The signature, with timestamp and optional photo, becomes part of the immutable case record.
The hotel never has to argue later that the item was returned. The receipt is right there in the file.