Genesys Cloud CX

 View Only
Discussion Thread View
  • 1.  New Features Coming to STA on Dec 14, 2022

    Posted 12-14-2022 11:59

    Today we're really excited to be releasing some new great capabilities as part of Speech and Text Analytics

    Usability improvements to messaging and email in the Transcript tab
    Genesys Cloud improved usability for messaging and email in the Transcript tab of an interaction's overview page. Chat and messaging interactions on the Transcript tab now include dates and timestamps to see when the message occurred and different colors to distinguish between internal and external users. Users can also now toggle HTML content for all emails in a thread instead of for each email.

    Improvements to Topic Spotting
    It's now possible to add words and phrases in quotes to ensure that they are mandatory during topic detection. This ensures that Topic Spotting is even more accurate, picking up only those phrases that match the specific requirements detailed in a Topic.

    Customer Sentiment in Agent Queue Detail and Queue Agent Detail
    We've added sentiment data into the Agents Queues Detail and Queue Agent Detail views, giving supervisors more insight into customer sentiment in these views.

    Support for Japanese (ja-JP) Topic Spotting in Voice Transcripts
    We've added support for Japanese (ja-JP) for Topic Spotting in Voice Transcripts, allowing more customers to take advantage of the capabilities of Topic Spotting.

    These capabilities further increase the power of Speech & Text Analytics in Genesys Cloud.

    For more information check out the release notes:

    STA Team

    Anik Dey
    Genesys - Employees

  • 2.  RE: New Features Coming to STA on Dec 14, 2022

    Posted 12-15-2022 11:24
    Edited by Peter Stoltenberg 12-15-2022 11:28
    Hi Anik,

    On the new ability to add mandatory phrases, it looks like there is a typo in the documentation "this means phrase order is important important".. Maybe it's trying to say "more important? I was looking, so figured I'd let you know :)

    And as I'm trying to properly understand this.. Is it fair to think about it in the following way:

    1. If I have my strictness set to 60%, the S&T algorthim is going to try to match the detected utterance to the topic. If we pass the 60% by matching a pre-defined phrase,(say "cancel my credit card"), where credit card is the mandatory word, then we can move onto the next step.
    2. Then, once we pass the strictness level, we look to see if there are any mandatory words, and if there are mandatory words in the phrase, we check the utterance to see if it contains that word (or multiple words).

    Does it then make sure that the mandatory word is in the right order in the utterance, in comparison to the defined phrase? Or, how does that work?

    Peter Stoltenberg
    Avtex Solutions, LLC

  • 3.  RE: New Features Coming to STA on Dec 14, 2022

    Posted 12-15-2022 11:53
    Hi Peter

    Thanks for spotting that, I'll get that changed asap!

    Generally what you've said above makes sense. We'll generally look for phrases as we do today, but in order for them to be a 'hit' against a topic they must contain the mandatory words/phrases.

    It's a slightly simplified example, but if we had a phrase in a topic "want to "cancel" my contract" but found "want to extend my contract" in the transcript, we would not have identified that topic, as "cancel" MUST be in the phrase for it to be detected.

    Hope this helps


    Andrew Boland
    Genesys - Employees