What’s new in Voicify so far in 2022

by | Mar 14, 2022 | Voicify News | 0 comments

2022 is off to a great start at Voicify, and the team has been busy adding many new features that make it easier than ever to build, deploy, and manage conversational voice assistants. Authenticated Voicify platform users can access complete release notes in the Voicify Knowledge Center, but we would like to share some details about a couple of new features.  

“Try Voicify” version 1.1  

Anyone who has ever worked in a content management role knows how hard it can be to get busy executives to log into a technology platform to approve something. With the initial release of “Try Voicify,” workflows around testing and approval became more effortless than ever by making it possible for anyone with a web URL to test a voice assistant created with Voicify. Whether they have access to the Voicify platform or not, “Try Voicify” allows users to experience a Voicify custom assistant quickly and easily through a web browser.  

The initial release of “Try Voicify” relied on browser text-to-speech. This approach worked for baseline functionality, but we knew it could be improved because a browser’s default voice is unnatural and does not support SSML.   

Now, “Try Voicify” leverages Voicify’s new SSML to speech endpoints found in our assistant API. With it, we support SSML, including the <audio> tag that allows conversation designers to include custom voiceovers and all the other SSML tricks you can configure in Voicify.   

Speaking of customization, “Try Voicify” now also supports Markdown. This addition allows conversation designers to apply custom formatting and embedded links using standard markdown format.  

The latest release also includes improvements to the user interface. These include:  

  • Hints are now scrollable and render better.   
  • Improvements to image rendering.   
  • Improved mobile scaling  
  • Added a splash screen for text-to-speech  

Released in mid-February, version 1.1 includes a few other new features to improve functionality, efficiency, and customization options. Voicify customers can get a complete run-down of what’s included and a look at what’s ahead for “Try Voicify” in the release notes 

If you’re curious about the end-user experience, check out this demo created by Aadu Pirn, architect of the Voicify platform.    

Required Phrases and Confidence Scores  

With the the 3.5 release, a powerful feature is now available within the general user interface that was previously only available in the API: Required Phrases and Confidence Scores.   

Together, these two components allow the conversation designer to control how specific user inputs must be at a particular point in the conversation instead of the entire discussion.   

Chief Product Officer Alex Dunn describes the feature this way:   

“Required Phrases and Minimum Confidence Scores allow the conversation designer to manage when a user can surface a Question-and-Answer conversation item. A required phrase must be in the statement or question from the user and require an exact match. If, for example, the conversation includes a moment that allows users to ask about different foods, the designer can add other food variations as required phrases the user may say. If the user says something close but doesn’t contain one of the required phrases, they won’t hit a designated conversation item.  

The conversation designer can then use the minimum confidence score to handle how close the user’s input must match one of the input variations on the conversation item. This value can be any number from 0 to 1 where 0 does not add extra strictness, and 1 requires an exact match.”   

This feature dramatically improves discussions about finding something like a product, doctor, symptom, location, or any conversation that requires the user to be more specific. As an ancillary benefit, Voicify also leverages the data around required phrases and confidence scores in the machine learning model used to enhance clarity and accuracy across the voice assistant, which is particularly helpful in applications about specific knowledge domains.   

Would you like to know more?  

The team is working on many new features to make it easier to build, deploy, and manage custom assistants available to users everywhere. If you’d like to know more about any of the new features added to the Voicify platform or what we’re planning for the coming weeks and months, we’re happy to connect anytime. Please feel welcome to contact us for a no-cost/no-commitment conversation, or you can go ahead and use our online scheduling tool to book a demo right now.

Time to talk? We’d love to connect.