| Add-on | AI agents - Advanced |
Conversation flows allow for multiple interactions between the AI agent and the customer, offering greater automation potential. For example, let's say a customer emails a query, but leaves out key details, such as order ID. The AI agent can request the missing information before going any further.
Conversation flows for email notifies the AI agent when there is a new customer reply, and recognizes when a live agent is needed. Using these data points, you can track reopen rates and measure automation performance for the AI agent.
This article covers the following topics:
About new ticket node types in conversation design for the EAP
The early access program includes new node types for ticket AI agents in the dialogue builder. This allows you to build dialogues with multiple AI agent replies, including additional replies from your customer and decision points for escalating to a human agent.
The following new node types are in the dialogue builder for ticket AI agents:
-
Visitor message, with the following children node configurations:
- Intent predicted: Use if you want to allow the customer to trigger another intent, following it with a link to reply node.
- Entity recognized: Use when expecting a content entity in the response.
- Free text written: Use as a fallback or follow with a conditional node to analyze the content of the reply.
- Link to reply
- Multiple ticket replies
Understanding conversation flows for email scenarios
Conversation flows for email supports automating the first triggered intent in every
conversation. When the dialogue of the triggered intent is complete, and the ticket will not
receive more responses from the advanced AI agent, the tag
escalated_by_ultimate is added to the ticket to indicate that the
automation is complete.
escalated_by_ultimate tag is applied in the following scenarios:
Using single-reply dialogues
Single-reply dialogues with conversation flows work the same way as without conversation flows. If nothing follows a ticket reply node in the dialogue builder, the automation is considered complete.
Using link to reply
While conversation flows for email is limited to one intent per conversation, linking allows you to continue the conversation in another dialogue, so the link to reply node is not considered the end of the conversation. Even though it's the final node in the dialogue, you can link to other single- or multi-reply dialogues.
Using multiple-reply dialogues
Multiple-reply dialogues can use the additional block options available for conversation flows for email, including visitor message, link to reply blocks, and multiple ticket replies.
These dialogues allow the AI agent to send follow-up questions to confirm customer intent or request missing information. Additional messages from the visitor can be processed and used during the automation.
The number of blocks is not limited, but it's a good idea to use no more than three messages to gather the required information to automate a case.
There following events are supported for advanced AI agents for email, and can be triggered in multiple-reply scenarios:
- Ticket received by AI agents - Advanced: This event is triggered only once per conversation, when the first customer message is received.
- Ticket processed by AI agents - Advanced: This event and the related actions are triggered with each ticket reply.
- Ticket reply delay events: The delay events are triggered for each reply. When using delays with conversation flows for email, don't set long reply delays, but consider reply delays of 10-30 minutes.
Escalating from a multiple-reply dialogue
When escalation is needed in a multiple-reply scenario, following the visitor message or conditional node, you can add an empty ticket reply node. The empty reply adds the escalation tag to the ticket, can trigger additional actions, and does not send a reply to your customer so a live agent can take over.
Limitations for the conversation flows for email EAP
The conversation flows for email has the following limitations:
- Automation is limited to one intent per conversation
- No rule-based escalations
- No AI agent replies to tickets older than three days
- New analytics metrics are not yet available
- Email conversations don't support tickets created through the Zendesk API
Allowing advanced AI agents to send multiple replies
When conversation flows for email is on, your advanced AI agents add the
escalated_by_ultimate tag to all tickets processed by the AI agent.
Advanced AI agents rely on that ticket tag to know when to provide a response.
To set up your advanced AI agent to send multiple replies, you need to first update the AI agents - Advanced automation trigger. Then you can build a dialogue with multiple ticket reply nodes.
- If your advanced AI agent is already live, wait at least three days from turning on conversation flows before you update the automation trigger. This gives the AI agent time to mark all current conversations as complete, so that if a ticket is reopened, it won’t be answered by the AI agent. This is a known limitation of the EAP that AI agents do not reply to tickets that are older than three days.
- If you are still building your advanced AI agent, for the first 72 hours, modify your trigger so that Ticket Is Created is the only condition in the ANY section. After 72 hours, you can re-add the Ticket Is Updated condition to the trigger.
If you have a new advanced AI agent or haven't yet modified the automation trigger, you can simply remove and re-add the trigger in the Settings > CRM Integration > Integration menu.
- In Admin Center, click
Objects and rules in the sidebar, then select Business rules >
Triggers. - Open the AI agents - Advanced Ticket Automation trigger for editing.
- In Conditions, remove the Ticket > Ticket is Created condition.
- Add the following conditions:
- Ticket > Comment is Public: This ensures that the AI agent is notified of each customer, agent, or AI agent reply.
-
Ticket details > Current user is not YOUR AI AGENT: This avoids loops. Because
the AI agent now listens to all replies, you need to exclude the AI agent's replies so it doesn't reply to itself.
- In Actions, you need to update the object’s properties by replacing the JSON
placeholder with the following code:
{ "botId": "YOUR_BOT_ID", "message": "**{{ticket.title}}** \n\n {{ticket.latest_comment_rich}}", "conversationId": "{{ticket.id}}", "type": "message_multitouch", "tags": "{{ticket.tags}}", "requester": "{{ticket.requester.id}}", "currentUser": "{{current_user.id}}", "createdAt": "{{ticket.created_at_with_timestamp}}" }Changes to the JSON body include the new requester, currentUser, and createdAt properties. These changes allow the advanced AI agent to differentiate between agent and customer replies and to not reply to tickets older than three days. The type is also changed for analytical purposes.
- Click Save.
0 comments