Plan requirement
| Subscription | Suite Team or higher, Support Team or higher |
| Access | Admin |
Yes. An AI agent on email answers incoming tickets from your articles, in the same way as on messaging, and generative replies cover 100 languages and more. Connecting the channel creates a trigger that is set by default not to answer tickets submitted by agents, which affects your first test.
What happens when I connect the channel?
A ticket trigger is created for that agent, called Ultimate.ai Multitouch Ticket Automation followed by its ID. It becomes active when you publish, and watches public comments on tickets created or updated on that channel. Agent-submitted tickets are ignored by default, so a test from your own address produces nothing. It is not a condition but a field in the trigger's JSON: submitterRole and authorRole.
One thing does not carry across: an email AI agent cannot use dialogues. Anything you scripted for messaging stays there, and on email the flow comes from generative procedures instead.
How do I keep it safe?
Scope it. Use triage topics or tags so it answers only the subjects you are confident about, and everything else reaches a person untouched. An AI agent answering all incoming email is the configuration most likely to be switched off within a fortnight.
Email is riskier than messaging because there is no clarification: the customer wrote what they wrote and has gone, so a misunderstanding arrives as your considered answer, in writing.
Say in the reply that it is automated, with a clear route to a person, and read every email the agent sent in the first week. It is a finite number at the start, and it is the only way to catch the answers that were confidently wrong.
Comments
0 comments
Article is closed for comments.