by Sheana O’Sullivan
We have entered an age where biology and technology have intersected. Thanks to AI, Deep Learning software can imitate activities in the neuro-cortex of our brains. The software is continuously gaining more knowledge to recognize patterns in digital representations of images, sounds, and other information. And as a result, AI is now even better at recognizing voice.
Early Days for Enterprise
The sale of 50 million smart speakers in US homes shows us how widely voice has been adopted as a user interface. But the feeling amongst many US tech firms is that the use of voice today for enterprise is still fairly rudimentary and instructional. One of the largest tech companies in the Valley admitted to me that voice has not yet been used to understand and solve business problems. But, they added, this is now changing as voice becomes easier to use with the addition of transcription and AI functionalities.
At the recent Dreamforce event in San Francisco, Marc Benioff announced Salesforce users can now speak to the CRM, adding meeting notes by speaking into the mobile app. Their Einstein voice assistant then uses natural language processing to interpret what was said and automatically finds the account in Salesforce to add actions. Call transcription and audio recording are also now available in CRM to provide a 360 degree of customers.
Elsewhere, Microsoft has linked Cortana virtual assistant to Teams, its office chat system, so that users can start a video conference with a voice command. And I just read Facebook is voice-enabling its Messenger app.
Sentiment as a Service
But there’s so much more to voice. Voice is a real source of data. And a rich source of data, too. Voice provides insight into emotion, personality, intent and sentiment.
In customer service, one of the leading FinTech companies is using voice for customer analysis. Sentiment analysis maps the acoustic characteristics of a customer’s voice, like stress, cadence and speed, to the context of the conversation, producing a single score. This call score can be used to measure relative sentiment or emotion across various cross sections of calls, agent groups, and time frames and, ultimately, improve the customer experience.
Companies are gaining competitive advantage by listening to the voice of the customer to uncover customer attitudes on services, products and, just as importantly, how customers feel they are being treated.
In the medical field, analysis of the voice (phoniatrics) is being used to identify dementia in elderly patients. A few weeks ago, I met with a Danish start-up that’s using voice and pattern recognition technology to predict cardiac arrest in real time, enabling first-aid responders to arrive faster and provide better-informed treatment.
I recently did some work for a US start-up that has developed a biotechnology platform to analyze the voice and deliver neurofeedback. Their proprietary algorithm analyzes the unique frequencies of the voice to provide insights into well-being. I actually was given the chance to test the pre-release software, too. The report highlighted tightness in my body which I had been caring for and a vitamin deficiency that was confirmed by my doctor.
Challenges and Opportunities
But like all other forms of data, privacy and security pose a significant challenge. Some solutions on the market today help mitigate this risk by enabling processing to be done on the device, protecting voice data and never sending it to the cloud. We are also seeing voice biometrics recognition for user authentication and security.
Another challenge is how accurately machines interpret the ambiguities in language – there is a heavy dependence on context in human conversation. But with advances in NLP, machines are learning to respond more intelligently and will be able to perform more complex tasks during a dialogue. Soon, we’ll be able to chat back and forth to our devices. I look forward to that – that will make a great night in with Alexa!
As people start to engage more with smart speakers, and their AI capabilities grow, the potential for voice will only become clearer. Over the next five years, I think we are going to see voice technology transform not just business and computing, but the quality of our lives and well-being.