Configure support
Access the support settings
Go to Agent Settings and select the Support section. Here you can define what your agent will do when it encounters a case it can’t resolve on its own.
What the agent does when it can’t resolve a case
This is the key question in the support settings. You have four options:
Nothing - Respond as best it can
The agent will try to give the best possible answer with the information it has, without taking any special action. It won’t create tickets or transfer the call. This is the most basic option and can be useful if you prefer to handle difficult cases yourself by reviewing conversations.
Flag - Mark as pending
The agent will log the case as a pending support ticket and collect the necessary information from the customer. The ticket will appear in your Support section for your team to review later. This is ideal if you want the agent to take note of the problem and your team to handle it afterward.
Escalate - Transfer to a human
The agent will transfer the call to a phone number you configure, so a real person can assist the customer directly. Before the transfer, the agent can provide a brief summary of the case so the person receiving the call has context. This option is only available for calls.
Decide - Let the agent decide
The agent will use its judgment to choose the best action for each case. Depending on the situation, it may respond as best it can, flag the case as pending, or escalate to a human. This is the most flexible option and gives the agent autonomy to adapt to each conversation.
The Escalate option (transfer to a human) is only available for the calls channel. If your agent works exclusively with messages, use the Flag or Decide options.
When to use each option
- Nothing: businesses that prefer to review conversations manually and don’t need a ticket system.
- Flag: teams that want an organized record of pending cases for follow-up.
- Escalate: businesses with staff available to take calls in real time when the agent can’t resolve the issue.
- Decide: when you trust the agent to make the most appropriate decision based on the context of each conversation.
