Technology
World-leading understanding of language driven by deep learning

Myra uses deep learning and other cutting-edge techniques to read customer issues like a human would. These techniques give Myra powerful language recognition abilities that offer high performance while minimizing the amount of training and preparation needed to get started.

Myra's models have been trained on over a billion pieces of web content, giving them baseline knowledge of a vast number of words and language concepts. Myra uses this foundational understanding to provide immediate language accuracy far above that of previous generation language understanding systems.

Automatic understanding of word relationships

No need to train Myra on the thousands of different ways customers might state their issues. Train on “taxi”? Myra will automatically recognize “cab.” Our internal mapping of word meanings give Myra continual support in understanding the extreme variety found in real-world language.

Myra's internal word relationships provide additional accuracy. For example, without additional training, Myra knows that 'bmw' and 'audi' are closely related concepts.

Incorporating syntax's effect on meaning

Many words have multiple meanings. For meaning to be determined, these words have to be placed in context. This confounded many previous natural language understanding systems. For instance, depending on the context, "Queen" could mean Queen Elizabeth II or the rock band. Myra reads sentences from left to right, looking at not only each word, but how it is used in the sentence, boosting accuracy.

Subscribe for the latest!

Don't miss news on conversational AI, automation, and our latest announcements