The band is not a wallet

Two architectures, one of which is wrong

A cashless wristband can either carry an identifier that maps to a balance held on your server, or carry the balance itself in the chip's memory. The first is how nearly every well-run programme works. The second is proposed periodically, usually on the grounds that it keeps working when the network is down, and it introduces a failure mode that the first one does not have.

A damaged band should not be lost money

If the balance lives on the server, a broken or lost band is a credential to disable and reissue — exactly like a lost room key, and the guest's money is untouched. If the balance lives on the chip, a band that fails has destroyed value that may or may not be recoverable, and now the property is in a dispute about how much was on it. There is no good way to end that conversation, and it will happen in front of other guests.

Offline is a real requirement with a better answer

The argument for stored value is genuine: beach bars sit at the end of the longest cable run and the wireless link fails on the busiest day. The right response is an offline mode at the till that queues transactions and posts them when the link returns, with a decision made in advance about whether queued spend can overdraw a pre-paid balance. That handles the outage without moving the money onto a piece of silicone.

Ask where the money lives

It is a single question and it separates a considered proposal from a sales pitch. If a supplier cannot answer clearly, or answers that the balance is stored on the band because it is more secure, that is worth pursuing. Security is a property of the system and the process around it, not of where a number happens to be written.