Virtual Service Agent - Default text and handling opening requests (escalations)

[Robinson_ Andy]

Currently exploring the Virtual Service Agent, but there's a couple concerns I don't see any obvious way to change, so I wonder if I'm missing something or if anyone has a way to handle it. 

You can of course create as many intents as possible to try and cover possible questions and if you've got a clean Knowledge Base, the AI answers might be of some use. 

The default behaviour opens a request if it hasn't matched an intent and the user says the AI answer didn't resolve their issue.

The problem with this, it doesn't display the text from the standard flow 'Escalate', which by default is 

I’ve gone ahead and raised a request for you - a human agent will get back to you soon.

You can update that which is fine for your intents. However testing it, when an AI answer hasn't resolved the issue it automatically creates an issue as your default issue type, but with the following text:

PROJECT-JIRA-ID is your request.

Firstly I'd like the option not to create an issue, as the issue is largely meaningless. It's just their conversation dumped into a Jira, if your normal portal has requirements for opening a Jira - basic stuff like explaining in detail what you require, this doesn't capture that. 

If that's not possible, I'd at least like to be able to change the text "is your request." to inform the user they should leave a comment explaining what they need in further detail. 

While the intents are limited too - you can't use their input for many field types - you can at least ask for information before creating the Jira.

I suspect the answer right now is to just increase the intents to try and cover the questions that cause these Jiras - any ideas?

[Rob Mkrtchian _CAIAT_US_]

You can try this setup within the intent. When user confirms that his request was not resolved, you can add steps and ask for more information and then link to the fields that you want.

Second, if you don't want it to create a request just change the last step to "resolve" instead of "escalate".

I have attached a screenshot to show how to do it. Let me know if that helps.

[Robinson_ Andy]

Yeah for intents it's fine, as even though you can't do a lot with custom fields (yet), you can at least prompt for more information. 

It's outside of the intents which is the problem, if it doesn't match any intents and the user says it hasn't resolved their query, or it attempts an AI answer based on the knowledge base, the user can still click no. 

When the user does that it never goes to an intent, and by default it creates a request as whatever your default is set to. 

---

My current workaround is to set that default to a hidden request type, with its own issue type - both called something like JSM Virtual Agent. That way in any intent I can change the request type before an escalation as normal. 

The crucial part is, there's then Jira Automation that picks up any requests opened as this JSM Virtual Agent escalation, closes it as 'WON'T PURSUE' and adds a comment informing the user it's a new process and this will be used to further train the agent and links to self-service documentation along with a link to the JSM Portal to raise a request manually. 

It's not as smooth as it could be, but it seems the only way forward that doesn't create a lot of Jiras that cannot really be actioned.