Last July, Google announced its Contact Center AI product for helping businesses get more value out of their contact centers. Contact Center AI uses a mix of Google’s machine learning-powered tools to help build virtual agents and help human agents as they do their job. Today, the company is launching several updates to this product that will, among other things, bring improved speech recognition features to the product.
As Google notes, its automated speech recognition service gets to very high accuracy rates, even on the kind of noisy phone lines that many customers use to complain about their latest unplanned online purchase. To improve these numbers, Google is now launching a feature called “Auto Speech Adaptation in Dialogflow,” (with Dialogflow being Google’s tool for building conversational experiences). With this, the speech recognition tools are able to take into account the context of the conversation and hence improve their accuracy by about 40%, according to Google.
In addition, Google is launching a new phone model for understanding short utterances, which is now about 15% more accurate for U.S. English, as well as a number of other updates that improve transcription accuracy, make the training process easier and allow for endless audio streaming to the Cloud Speech-to-Text API, which previously had a five-minute limit.
If you want to, you also can now natively download MP3s of the audio (and then burn them to CDs, I guess).