Announcing the Summarization & Text Enhancing EAP

Exciting news! Great complement to Intelligence in the Context Panel.

Is the data shared on those kept secure and not used by their GPT engine for its training?

Amazing feature!

Another benefit for the agents could be to use ChatGPT to parse issue description and suggest steps to resolve based on knowledge base articles. Do you have anything like this on a roadmap?

Will this functionality be made available via API? It would be useful to be able to make use of the ticket summaries in custom functionality we have built internally. 

Can this be added to a Sandbox account for testing and then the live account later if it works well?

Hey @rafael20  thanks! OpenAI does not use any data coming from Zendesk to train their models.

Hey @anton14 thank you! There's something like this for Chat: https://support.zendesk.com/hc/en-us/articles/4408843471642-Enabling-Article-Recommendations-in-Web-Widget-Classic- that does not use ChatGPT. Also, we have a few things in our roadmap that touch on the subject - expect more news this year! 

@zachary17 for now we have no such plans, but I understand your use case. 

@kyle43 please complete the EAP form with your sandbox information and we'll see what can we do.

Thanks @jakub14, my question has since been answered via the Zendesk AI Data Use Information article.

How is the language detection within the Intelligence section affect workflows or other items?  Is this separate from the native Zendesk Language detection?

We have started using this feature and would love the ability to undo an enhancement we have added via tone shift. 

Is there no way to see if the agent used the "Enhance Writing" feature to modify a reply? 

@jakub14 Curious if you can answer my question above: is there a way to see if an agent used the "enhance writing" feature on their reply? Is that something tracked within the ticket, the same way we can see that agents applied a macro, for example?

Also, will there be a way to fine-tune the "voice" used for the "enhance writing" features? For example, the "make more friendly" feature can be helpful but is a bit too informal at times (at least for our brand voice). If we could have it, say, scale back the exclamation points, and make other adjustments or even provide instant feedback on the result, that would help. 

I came here to ask the same question as @michael28. This will help, especially when our leads are reviewing the interaction for quality assurance, to know if it was the tool or if the agent wrote the communication. 

Hello,

Will there be a way to customize summarization promts?

In our operation some teams could make use of a more bullet point format rather than a paragraph and other team might make use of a more intent-based summary.

So the idea is to have a setting in admin center, to assign different summary prompts to different contextual workspaces.

CS group can have a prompt saying; "summarize this content by making bullet points." and a sales group can have a prompt saying "summarize this content focusing on personal statements."

This would allow different teams to have more tailor-made summaries related directly to what they need to see.

Love this! Been testing out the 'expand' feature and it's great. 

I noticed the AI uses "Hello [name]," but we usually use "Hi [name]" (hi and no comma). Is there going to be a way to standardize this somewhere so that the AI generates our preferred greeting?

Thanks!