Curo® Speech and Language Platform
The new wave of speech, language and multimodal interface solutions
Available today: the technology
for tomorrow's solutions
The devices, applications and services of tomorrow need the latest foundation. That's why Interactions offers speech, language and multimodal interface solutions that deliver unprecedented accuracy and performance.
With our Curo platform, you can take full advantage of automatic speech recognition, text to speech, voice biometrics and natural language processing. Years of applied research in machine learning and deep neural networks positions our technology at the forefront of the industry. Our service-oriented architecture provides a broad set of speech and language platform APIs for partner applications and third-party developers. All available via the cloud or on-premises.
Automatic Speech Recognition
Deliver natural and successful communication experiences for your customers.Learn More
Text to Speech
Give voice to your content.Learn More
The future: Decades in the making
In 2014, Interactions acquired the AT&T WatsonSM technology platform that's shaped the industry for decades. Now that we've integrated it into our own patented technology, Interactions can deliver even more accurate and conversational speech and multimodal solutions.
Here are just a few highlights showing the Watson platform's history of innovation:
2015 and beyond
Seamless communications, information access, personalized commerce and customer care are made possible with intelligent speech and language multimodal services. This includes technology to enable seamless, efficient communications for home automation, connected cars, entertainment services and consumer interaction for sales, service and support.
Interactions acquires AT&T Watson speech and natural language platform to enable the next generation of efficient, multimodal applications, devices and services.
Interactions' licensing partnership with AT&T integrates Watson speech technology into the company's award winning Adaptive UnderstandingTM platform.
Interactions enters market with first fully conversational Virtual Assistant solutions that allow consumers to speak naturally.
In 2002, the more sophisticated "How May I Help You?” service went live, handling two million calls per month.
A paradigm shift occurs in user experience and automation with the introduction of natural language offerings. "How May I Help You?" conversational technology for automated customer care is first introduced. In recent years, AT&T has increased the flexibility of speech recognizers through dynamic hierarchal statistical language modeling, a strategy that supports recognition of complex grammars. New algorithms based on deep neural nets have been added to more accurately model the sound of the human voice and reduce error rates. A wide range of software and process inventions has also been developed to increase speech recognition speed and precision.
Billions of dollars were saved when Large-Vocabulary Speech Services were introduced. The initial operator services system, Voice Recognition Call Processing (VRCP), was deployed in 1992 and saved hundreds of millions of dollars annually.
AT&T Labs pioneered machine learning and neural net technology through the 80s and 90s and, in 1996 deployed a check reader based on neural nets that processed 15% of all checks written in the United States. Also in the 90s, other AT&T Labs innovations such as finite state machine decoders and discriminative training provided performance gains, especially for large vocabularies. Today, we use these technologies to deliver significant accuracy improvements in real-time speech recognition.
AT&T pioneered the speech service industry by introducing the "Command & Control" approach. Simple commands like "Press or Say" dominated the customer service experience. In addition, barge-in capability was introduced, which allowed callers to interrupt the application and further expedite the customer service process.
In the beginning
Early stages of speech recognition shows promise for human-machine communication. Simple phrases are recognized and AT&T introduces multi-speaker capabilities enabling recognition of more complex interactions.