Machine learning, artificial intelligence, and the continued evolution of speech recognition algorithms are advancing the state of intelligent virtual assistants (IVAs) and will continue to do so into 2017 and beyond, according to industry experts.
Opus Research has forecast at least $500 million in IVA deployments in 2017, a figure that’s expected to double by the end of the decade.
Fellow research firm Visiongain, meanwhile, has valued the global intelligent virtual assistant market at $1.14 billion.
Unlike most technologies, which tend to start in the enterprise and then go to consumers, virtual assistants used in business can thank their growth to the popularity and marketing behind consumer adoption of the technology, most notably Apple’s Siri, Amazon’s Alexa, and Microsoft’s Cortana.
But the advances in IVAs go far beyond those best-known examples. Interactions, IBM, MindMeld, Aspect Software, Next IT, Samsung, Nuance Communications, 24, and others have entered the business either directly or through collaboration with others. Microsoft even launched a 5,000-employee division devoted entirely to conversational technologies.
"2016 was an interesting year. Artificial intelligence came into play again as something that could help enterprises," says David Thomson, vice president of speech research at Interactions. "Companies were looking at how to apply artificial intelligence technology to their business strategy and how to use it to solve business problems.”
Thomson credits the growth in IVAs to three things: Speech recognition has become more accurate, doubling its accuracy in the past four years; algorithms have improved; and conversational speech has replaced structured, static interactive voice response systems that could understand only specified commands.
As a result of those advances, IVAs continue to get introduced into a number of environments, from contact centers to hotels to financial services to personal services, with more advanced uses expected as the underlying software continues to evolve. Among functions under development is the ability for IVAs to listen to multiple people at the same time and to process each input separately.
Advances in the Call Center
Today, calls to an increasing number of contact centers will result in systems asking about the callers’ needs, with IVAs telling callers that systems can handle most requests. Rather than offering the caller an IVR menu tree, the IVA asks the caller to describe in a few words what he needs. As IVAs evolve, fewer queries will have to result in automatic defaults to live agents.
According to Amit Ben, chief technology officer and founder of Nanorep, deploying a web-based IVA in the call center of tax and bookkeeping software firm Inuit reduced the number of calls going to live agents by as much as 45 percent.
“You’re seeing an explosion in machine learning and machine intelligence that has been driving improvements in speech recognition during the last three years,” Ben says.
There’s increased interest in using IVAs to handle inbound call requests, agrees Rick Collins, president of enterprise at Next IT. “We’ve helped many of our target customers put [IVA] strategies in place. There’s an increased creation of bots and task-oriented systems for financial services and for travel.”
Business are using IVAs to speed up the customer journey, facilitating purchase decisions regardless of channel, says Jennifer Snell, Next IT’s vice president for enterprise. “There’s a big push for a seamless omnichannel experience, so companies are trying to figure out how to provide seamless conversational interfaces.”
Successful IVAs should be able to offer customer assistance through data entry and through speech, experts agree. Speech communication might be easier for a user in some instances but a less helpful option in others, such as extremely noisy environments or when a customer needs privacy (e.g., when having to disclose personal financial information).
Further advances in IVAs for home and business uses are expected as the underlying technology continues to advance and the demand continues to grow.