Trends in Speech Technologies for 2019

In 2018, we saw speech technologies rapidly evolving and becoming an even more natural part of our everyday life. But what’s going to happen in 2019? What trends will appear within the speech technology industry? We asked three influential Phonexia opinion leaders and got eight predictions for 2019.

Petr Schwarz, CTO

Voice Assistants May Increase the Gap Between Countries
Voice Assistants are finding their way into almost every device possible. But this trend will create a bipolar world at the beginning because, for the manufacturers, it won’t be profitable to focus on less-frequently used languages. So, there will be countries where voice assistants are everywhere and, on the other hand, also countries where voice assistants are totally absent.

Call Centers Will Decrease Costs Using Robots
The potential of cost optimization for many call centers using a human workforce is almost exhausted. This is why voice assistants will be replacing human operators in call centers more and more, even if their capabilities at the beginning are quite limited.

Voice Assistants Will Penetrate the Manufacturing and Services Industries
The current practice when employees have to interrupt work just to enter notes into the internal information system will be replaced by a natural dialogue with a voice assistant during the work itself. This trend will be more visible especially in industries where employees have their hands full with something (for example, in logistic centers, production halls, hospitals etc.), and later will spread to other industries as well. This will dramatically increase the effectivity of work.

We Cannot Trust What We Hear or See
Artificial intelligence has come so far that humans cannot distinguish a real voice or picture from an artificial one anymore. To change the voice or face of a speaker in a video has never been so easy and available as it is now. Anybody can pretend to be somebody else, which will dramatically increase the spreading of fake news.

Michal Hrabi, CEO

Voice Morphing Is on the Rise
Contact centers will start using voice morphing so their agents get digitally enhanced voices, which is more pleasant for customers. Celebrity voice samples will be used as well, so we might have the feeling we’re speaking to Gorge Clooney or Jennifer Lopez.
At the same time, the same techniques will be used by fraudsters to hack services accessible over the phone. Sophisticated voice biometrics systems will be able to recognize such voice enhancements if they are in place as an additional level of security.

More Security with Voice Biometrics
More online and even classical services will use voice biometrics for the authentication of speakers—usually doing this in the background without the user even noticing it (passive enrollment). The user experience will be upgraded and security improved.

Voice-Enabled Robot Process Automation
Robot process automation (RPA) is the automation of processes that are repetitive by nature and performed in organizations, such as calling customers to remind them about the date of their next payment, filling-in various forms, etc.
RPA using a voice interface will be a strong improvement in the productivity and cycle time for many robotic tasks so far being performed by humans. Contact centers, in particular, would be able to automate a lot of work with a voice-enabled RPA, thereby solving the problem of hiring and training new personnel repeatedly.

Michal Kral, Head of Software Development

A Demand for Customized Language Models
Companies will require language models that reflect their real customer data and take into account their legislative needs. They will need results with maximal accuracy, and general models will not be enough for them. This is why language models will be customized more and more for particular cases without letting the data leave the company’s infrastructure.

These predictions for 2019 seem to confirm how commonplace speech technologies and especially voice biometrics is becoming. From call centers to providing sophisticated levels of security, this technology can deliver actionable data, insights and defense against fraud that were never before thought possible. Its ability to cross language barriers and reduce the risk of human error while seamlessly integrating and operating in the background means in the coming years, we can expect to see the use of voice biometrics extend into even more areas of industry and life.