System replies, unlike intent replies, do not answer queries, but instead deliver messages.
You should include system replies in your conversation design to give visitors the best user
experience. Reply text is suggested, but you can adjust the text for tone of voice, to include
some FAQs, or to define alternate escalation strategies.
The system replies include:
- Collect (BSAT) reply: Collects feedback from your customers to measure how your AI agent performs. This reply is sent at the end of each conversation, asking the customer to rate their support experience on a scale of 1 to 5.
- Default reply: Informs customers that their query is not understood. This reply is sent when the AI agent lacks confidence about which intent should be triggered. This is an important reply in that it can help the user move on and not get frustrated from being misunderstood.
- Escalation reply: Used to configure escalations to a human agent.
- Failed escalation reply: Informs customers when there are no agents to take the chat. This could mean that no agents are online or it could mean that the agents online are all busy.
- Technical error reply: Informs customers when there is a technical issue. This reply is sent when there is an error, either from the CRM or the advanced AI agent. For example, when an unexpected error occurs when connecting to an external API.
- uGPT reply: When UltimateGPT is activated and no intent is recognized, the AI agent will reply. It generates responses to visitor messages with the help of UltimateGPT.
- Unsupported language reply: Used when the user writes in an unsupported language. This reply is sent when language detection confidently identifies the user's language as one that is not active in the AI agent. In that case, the AI agent uses the default language. In the reply, consider including the languages the AI agent supports.
- Welcome reply: Greets customers when they reach out and sets the tone for the conversation.