Intelligent Adaptive Voice Assistants – The Future of Speech Recognition

Multilingual Voice Assistants

The U.S. Census Bureau recently released data that predicts that by 2050, 54 percent of the American population will be minorities. Given the continuing rise of mobile devices, we can also predict the majority of Americans will be connected through mobile technology and will begin hitting language barriers when using their devices to communicate with companies.

With such a significant percentage of potential future customers possessing accents with pronunciation variations, organizations should begin to invest in tools that can help support multiple language variances to communicate through speech recognition programs that will ensure a quality customer experience.

As a person of Indian descent, I’ve struggled with interactive voice response systems (IVR) not understanding my accent and I often have to resort to using text-to-speech or asking to be transferred to a representative. With voice enabled technologies increasingly used to identify users, codify, search, navigate, and control, it is imperative for speech recognition programs to be accurate and deliver consistent experience across accent variations. If a voice-based solution hopes to succeed, designing one that is responsive to a bigger demographic audience will be necessary in the coming years.

A potential solution to consider: developing a database of thousands of variations of key words that are known to be pronounced differently across accents that could be checked with voice biometrics. Every accent profile would have certain unique settings that would boost speech recognition accuracy for that accent and ultimately improve a caller’s experience, regardless of their background.

For example, an organization implementing IVR or speech-enabled mobile applications could implement design solutions that enable their mobile application to detect ongoing speech recognition issues, and adapt intelligently. This application solution would involve the customer repeating a few keywords once and immediately assigning an accent variation to the customer’s personal profile.  The platform would recognize the customer’s accent in every subsequent use. Speech recognition processes would improve, eliminating the annoying process of having to constantly repeat the command or give up and transfer to an operator.

For companies looking to expand into emerging markets such as India, Brazil or China, a robust adaptive speech recognition engine in your customer experience portfolio will give businesses a strategic advantage.

Roadblocks such as improving speech recognition solutions to adapt to customers with various accents will be a major factor in the voice industry in the near future. However, once solutions are in place, they will help businesses gain a strategic advantage and improve overall customer experience.

It’s time to partner with experts who are not only looking at solving problems facing the industry  today, but also have an eye on innovation opportunities for  the future – specialists who are willing to think out of the box and ahead of competition. Only by putting the customer first and creating innovative solutions can businesses gain a strategic advantage and improve overall customer experience

For more information on using voice biometrics to improve the customer experience, check out our white paper here.

Share

About the Author: Sharda Divekar is the Senior Consultant for Angel’s Professional Services (PS). Sharda has a strong background in Telecommunications, Sales Engineering and Network Management. Prior to Angel, she was involved in various performance optimization, project management and client management assignments across global wireless carriers and infrastructure providers. As a part of the Professional Services team, she assists clients with identifying Caller First opportunities, as well as implementing and driving continuous improvement from Caller First IVR's. By leveraging her technical background and marketing insight, she likes to blog on experience-driven IVR design opportunities. Sharda has an M.B.A. from University of Maryland-College Park, and M.S. in Telecommunication from University of South Florida.

One Response to “Intelligent Adaptive Voice Assistants – The Future of Speech Recognition”

  1. Great post! Your vision is spot on with where the industry is headed. Case in point, Samsung S4′s key differentiating feature: “S Translator” -
    No more language barriers!

Leave a Reply