What's new

Natural Language Address Capture sub dialog module for IVR

[24]7 recently launched Natural Language Address Capture sub dialog module for IVR, which provides a seamless, natural caller experience and enables the caller to speak the entire address in one utterance.

Natural Language Address Capture intelligently validates the address based on a proprietary internal service. If the caller omits the zip code or provides an incorrect zip code for a valid address, we can append the missing zip code or fix the incorrect zip code. In the event that the address is not validated in a single utterance, the [24]7 Natural Language Address Capture provides a failback method-using the traditional three-step collection-to ensure the caller's needs are met.

For pricing information, please contact your account manager.

Deep Neural Network (DNN) Infrastructure

[24]7 recently deployed the new Deep Neural Network (DNN) Infrastructure for improved speech recognition in IVR. DNN guarantees better recognition results both for native and non-native speakers.

Why DNN?
  • First Enterprise Telephony Deployment Of DNN
    • [24]7 has exclusive access to Microsoft DNN technology
    • Speech models are tuned on 10 billion utterances per year (Xbox, Bing, and Cortana/Windows phone consumers)
    • Significant technology shift after 30 years of a GMM/HMM world
  • Digital Ready
    • One DNN model works for both wideband (mobile, web) and narrowband (telephony) applications
    • [24]7 provides REST API and VXML access to Speech and NLU services
  • Global Ready
    • EMEA and APAC languages being deployed in Europe, Australia, and New Zealand
    • 30% improvement in recognition accuracy for EMEA languages

For pricing information, please contact your account manager.


The primary goal of a voice application is to enable enterprises to reduce costs while offering a superior caller experience. Businesses strive to deploy applications that allow callers to complete their tasks efficiently and effectively without engaging expensive call center agents. This document describes a mechanism for measuring task completion within your voice applications in three steps. Learn how!

Using revisioning technology, Tellme Networks exposes a mechanism for you to revise and test your voice applications against modified or new functionality while simultaneously continuing to run your existing voice application on the production Tellme Voice Application Network. Once fully tested, you can deploy the revised version to your production environment. Learn more about Tellme Revisions!

Abstracting data from your VoiceXML code allows you to build customized applications that are easy to maintain. . Tellme's support for the proposed data element allows you to fetch data from the server without re-fetching the entire VoiceXML document. You manipulate the data in JavaScript using the Document Object Model, a W3C standard for programming XML documents.

Learn how to automatically share state between the VoiceXML code and server-side scripts in your application by using client-side cookies.

[24]7 Inc.| Terms of Service| Privacy Policy| General Disclaimers