Speech Continues to Advance in the Contact Center
On the IVR side, there’s been a rise in use of more conversational IVRs that act as a virtual agent.
“Speech and AI-driven IVRs can understand and respond to the caller in more complicated ways, such as answering without requiring many layered prompts or questions,” Aylarov says.
IVAs, meanwhile, are seeing adoption rates grow each year “because enterprise customers—pushed by their expectations with personal voice assistants—now expect voice technologies to feel fluid and convey conversational intelligence,” says Malingo, who cites Verint clients like Amtrak, the U.S. Army, and Novo Nordisk as increasingly relying on IVAs in their customer service stacks.
When there’s a surge in inbound queries to customer service centers, “speech-enabled IVAs can resolve simpler queries and can get the caller to the right agent quickly,” Moffat says.
When it comes to ASR, it can be either directed dialogue, where the application tells the caller which words or phrases will be accepted at a particular menu location, or natural language, which interacts more naturally with the caller (asking, for instance, “How can I help you?”).
“Directed dialogue has become quite common and successful with a relatively low implementation cost. Natural language typically requires additional speech science resources and expertise and can demand elevated budget considerations to implement but results in an incredibly fluid and conversational application,” says Bob Traynor, vice president of consulting services at Avtex, a customer experience solutions provider.
Another innovation garnering amplified attention is voice biometrics, which identifies a customer’s identity via active listening technology. Donna Fluss, president of DMG Consulting, says “voice biometrics can speed up call handle time and reduce fraud, although the technology can be expensive.”
Good Candidates
The businesses with the most to gain from boosting their customer service operations with speech have straightforward and repeatable navigation needs, where the speech required is concise and clear.
“Examples can include navigating a sequence of accounts or loans or simply routing to the most appropriate group of live agents,” Traynor says. “But businesses that have not communicated consistently with their customers or that have changed and adopted different words and phrases to describe certain tasks can have some challenges in implementing a sophisticated speech application.”
Enterprises that require highly personal and sensitive information, such as full Social Security numbers and addresses for identification, might not be an ideal fit, Traynor adds, “as there may be security and fraud concerns as well as low customer adoption.”
Speech-enabled customer care tools often work best for companies that implement these resources into their overall customer service ecosystem.
“The tools and technology are not seen or experienced in isolation, but rather as part of the overall journey,” says Carey Stoker, senior vice president of customer care services at Radial, a customer care and commerce technology provider. “The speech technology should have a persona that can be shared through a chatbot, an IVR, or guided experiences. Having a learning AI, speech-enabled system to provide insights into the actions and journeys customers are taking, as well as providing speech analytics on sentiment, intent streams, agent behaviors, customer experience, is the ideal scenario when it comes to customer service automation.”
Years ago, implementing speech technology in customer service used to be fairly difficult, because the systems required proprietary interfaces, but that has changed, too.
Now, solution providers are offering no-code tools for building call center platforms, which “make it much easier to integrate modern speech technology into customer service interactions, even for organizations without an IT leader to assign to the project,” Aylarov says.
That’s a benefit particularly appreciated in the wake of the coronavirus crisis.
“Due to COVID-19, many companies closed down their contact centers and converted their operation from direct call to voicemail,” says Derek Wang, founder and CEO of Stratifyd, an AI-powered analytics platform provider. “But speech technology has dramatically reduced the workload and costs for call center services.”
Roadblocks Ahead
Make no mistake: Leveraging voice as a customer service interface can present many challenges.
“Your voice tech needs to be sufficiently sophisticated to understand various caller accents and dialects and to sort out the customer voice amid lots of background noise and even a competing voice,” Sachdev says.