Problem statement:
Currently, when the AI cannot respond appropriately to a guest message (such as scheduling requests or damage reports), it sends a fallback message: "Got your message - we will get back to you shortly." If the guest continues to send messages, the AI keeps replying with the same fallback message repeatedly. This creates a poor guest experience and requires manual intervention to pause the AI in specific conversations.
Current workaround:
Hosts must manually pause the AI in specific conversations to stop repeated fallback messages. This requires active monitoring and intervention.
Impact:
- Guests receive repetitive, unhelpful messages
- Creates frustration for guests who are reporting issues or making requests
- Requires manual intervention from hosts to manage conversations
- Notifications are already sent to the team, but the AI continues to respond with fallbacks
Proposed solution:
Provide controls to limit when the AI sends fallback messages, such as:
- Restrict fallback responses to certain topics
- Limit fallback messages to being sent only once per conversation
- Allow hosts to configure fallback behavior settings
Unlock:
Hosts would be able to let the AI handle conversations more gracefully, with guests receiving a single fallback message before a human takes over. This would improve guest experience and reduce the need for manual intervention.