Agent Handoff
This sheet contains the list of steps when the bot should do an agent handover and the content to be displayed before the handover.
At various instances, an automated response by the bot might not be the best suitable option to address user queries. Hence, the query has to be handed over to a live agent. Also, it is suggested to send an acknowledgment message before the handover.
1. How does it work?
The sheet contains 2 mandatory headers — Intent and Text.
Upon providing the Intent and Text, the platform will be able to invoke agent handover workflow and display the acknowledgment message.
Below is the description and sample value for better understanding:
| Header | Type | Description | Sample |
|---|---|---|---|
| Intent | Mandatory | The intent is the objective/goal of the user that requires agent handover. It could be a single or multi-step goal. This should be selected from a pre-defined set of values. | track__order_delayed |
| Text | Mandatory | This is the message that needs to be displayed before agent handover. | We try our best to get your orders to you ASAP and we’re sorry to let you down. I’m reaching out to our customer service executive who will get back to you sooner. Thank you for your patience and apologies for the inconvenience caused! 🙏 |
In the above example, the bot will be able to understand when a user is frustrated since their order is delayed. It would first display the message “We try our best...”, and then hand over the query to a Live Agent.
NOTE: Utterances for every Intent is trained in the Core/Domain-specific AI models in the backend.
2. Sample Data
Below is a screenshot of a sample System Configs sheet:
3. Reference
Agent handoff for delayed order query:
Agent handoff for cash back query:
4. Best Practices
- Intent column cannot contain custom values.
- Text column can contain characters and emojis. Images and other media types are not supported.
- Do not add additional columns since they would not be recognized by the platform.
- If there is any request for a new configuration (eg: Agent handoff for Redeeming Points), then raise a request along with all the details. Product team would then go through the requirement and take the necessary action.
caution
⚠️ This is the link to the list of all available intents, suggested text, along with descriptions, and sample flows.