Provides speech resource building block components for on premise requirements, which makes recognition, semantic interpretation, and/or text-to-speech request via MRCP. The two primary elements of SpeechSuite are Nuance Recognizer and Nuance Vocalizer which can be used independently or together depending on deployment requirements. Additionally, advanced components are also available such as Dragon Voice for performing ontology-based semantic processing, Nuance Dialog Engine for IVR that runs IVR applications created with Nuance Mix tooling set, and Nuance Dialog Modules which are J2EE components that deliver a piece of packaged call flow in VoiceXML to accelerate the delivery for complex grammars such as names, addresses, email addresses, and spelling.

View Vocalizer Studio – Datasheet 

Nuance building block components are available as cloud-native containerized microservices for on premise or cloud deployments enabling the acceleration, development, and deployment of speech-to-text (STT) and text-to-speech (TTS) core technologies as well as Dialog (DLG) via a set of Nuance APIs. With these components, you will have the flexibility to build your own solution or if desired, engage in a collaborative approach with our highly experienced designers and professional services teams.

The next evolution of Nuance Recognizer ASR technology, NRaas understands and interprets spoken and touch tone input to help deliver seamless self-service. The underlying recognition engine is grammar-based, which means it requires grammars to interpret a caller’s utterance.

Constrained grammar recognition engines are especially useful for:

• Recognizing alphanumeric inputs such as account numbers and tracking numbers.
• Small scoped applications that are designed for interactions with a small set of words and phrases to recognize.
• Directed dialog to assist in navigating a menu tree of items, either used for the first interaction of a caller, or as the basis of the system’s conversational design.

The gRPC protocol in NRaaS allows an IVR application to request speech recognition in all the programming languages supported by gRPC that will streamline the migration of your existing NR-based IVR application to the cloud by decoupling your on-premise based SpeechSuite installation.

UniMRCP Connector is an on-premises software component that provides a native MRCP v1 or v2 bridge connection to leading providers of Conversational AI Services in the cloud such as Speech Recognition and Text-to-Speech.   The connector enables traditional on premise IVR’s using the MRCP protocol for ASR and TTS to now consume those resources in the cloud with limited modifications to your existing IVR application. 

ASR was made for contact centers. LumenVox scalable ASR technology can grow with your company while eliminating customer frustrations through enhanced accuracy and lowered call wait times. Start empowering your interactions with near-human converation. Transform your speech solutions with ASR, using NLP and NLU to interpret and grasp human language or convey information to your user with test-to-speech synthesis, with natural-sounding voices or effortlessly authenticate over remote channels, improving both customer experience and security through analyzing your customer’s voice pattern with LumenVox Voice Biometrics.

View LumenVox ASR – Datasheet 

View LumenVox TTS – Datasheet

View  LumenVox Supported Languages – Datasheet 

View LumenVox Voice Biometrics – Datasheet

Voicegain’s deep learning ASR with speech grammars or use of large vocabulary transaction for Voice Bots and IVR’s with support for eight languages all accessible via a standard MRCP protocol.

VoiceGain MRCP ASR – Datasheet 

Insights omnichannel reporting is a robust reporting and analytics framework that pulls data from multiple channels, formats it, and provides near real-time understanding of key metrics and changing usage trends to improve optimization and ROI. Custom KPIs can be set at the time the system is deployed. Drill-down and drill-through views provide detailed insight into application activity, enabling you to quickly identify specific opportunities for improvement.

Nuance Transcription Engine is a fast, highly accurate engine that produces search-optimized text output from an audio file. An accuracy of 80-88% can be achieved, which is one of the highest in the industry. Pre-recorded audio can be processed through NTE from 1x real time factor (RTF) meaning that it’s possible to process 60s of audio in 60s resulting in maximum accuracy, up to 10x RTF enabling processing of 60s of audio in just 6s. Nuance Transcription Engine can also transcribe live conversations with results returned incrementally, typically every 10-15s depending on user-defined settings. This means that transcription results are available approximately 10s after the audio is spoken.  The output format is a searchable lattice of time-stamped words or in a human-readable format.