etting things done – whether sending a message, calling Mom, checking the weather, buying diapers, playing your warrior music, or accessing the ship’s logs – all without having to lift a finger, is a sci-fi dream, now made reality with the power of voice AI.
In the last three years alone, sales of ‘smart home’ devices such as Amazon Echo and Google Home have surged exponentially. According to VoiceLabs, 1.7 million such devices were shipped in 2015. This was followed by 6.5 million in 2016, and an estimated 24.5 million in 2017 – totalling close to 33 million devices sold by the end of this year.
What makes voice AI all the rage now?
Since Apple’s Siri was first introduced in 2011, speech recognition capabilities have vastly improved, and now gone mainstream. Several speech APIs, are now readily available for developers to use. This, coupled with advances in AI and natural language processing (NLP), mean that voice interfaces will become increasingly prevalent, and provide significantly better user experiences. It’s no wonder we’re seeing a boom in voice search usage. According to Google, 20 percent of queries handled by Android phones were spoken. That’s 20 billion spoken queries every day. And, as more users hop on the voice bandwagon, the AI continues to learn, thus creating a virtuous cycle of positive user experiences, increased usage, and improved performance.
What exactly do we mean by voice AI?
You’ll hear the terms ‘voice AI’, ‘voice interface’, ‘virtual assistant’, and ‘chatbot’, spuriously thrown around, however not all voice AIs are made the same. Many apps today provide a semblance of intelligence through a voice interface e.g. you control it using your voice, or ask it a question, and it responds in a human-like way, sometimes imbued with ‘personality’. However, first generation chat assistants are notoriously riddled with limitations such as poor accuracy, inability to understand varying sentence structures and common slang, or shallow understanding – creating frustrating or weird user experiences, and limiting the interfaces’ usability.
Fortunately, voice AI is fast becoming more than just voice interfaces slapped on search engines, or machines acting on preset rule-based systems. Experts are continually working on improving machine learning and NLP capabilities.
The hope is that these ‘virtual assistants’ and ‘chatbots’ will be able to deliver better user experiences and function more as humans might e.g. to drive a car, or carry a conversation. And, in the not-too-distant future machines may not just think and learn like humans, but also to find new ways to discover answers to problems for themselves.
What impact is this having on businesses?
While voice AI is still in its relative infancy, we are already seeing it being used to enhance customer experiences, improve engagement, and streamline business operations.
Domino’s Pizza’s DRU Assist platform, for example, allows customers to order pizzas via various virtual assistant devices – without having to click through multiple pages of content, and typing address and payment details.
Health insurance provider, Humana, is working on providing a voice-based ‘MedBox’ solution to enable seniors to manage their conditions, by being able to ask voice assistants questions such as: “When is my next refill?” “What side-effects might I have?” or “I’m having nausea, which of my medicines might be causing this?” The company is also using emotion-sensing technology to simulate empathy for customers’ concerns and feelings to improve call center resolution rates and speed, and even allow home virtual assistant devices to sense when a customer may be showing signs of depression.
Aside from imbuing consumer products with voice AI, GE is putting AI into machines, easing maintenance reporting and reducing downtime in their factories, and helping customers like hospitals operate their scanners more optimally.
In the workplace, AI is already widely used as a productivity tool, scheduling lives according to their demonstrated preferences (e.g. FreeBusy Scheduling Assistant) , reducing the load of tasks such as answering HR and IT questions, and conducting interviews (e.g. JabaTalks), improving business intelligence (e.g. Sisense Eveywhere), and much more.
As the internet of things (IoT) becomes a reality, and voice search usage becomes the new norm, manufacturers will be putting conversational interfaces on everything. From commerce standpoint, businesses will then need to become more searchable and accessible via voice-based virtual assistants. Businesses will also stand to gain immediate value from voice AI implementations – whether by saving time and costs e.g. spent on menial tasks like CRM entry and transcription, or enabling real-time translation, or bringing in revenue e.g. by attracting and engaging customers with timely, relevant information.