By Yaniv Masjedi, Chief Marketing Officer at Nextiva
Agentic voices offer so much promise that businesses are moving quickly to adopt them. They can prevent huge losses from missed calls, answer common questions, capture leads, and free up human agents to focus on more complex tasks.
From insurance firms to legal services, healthcare, and hospitality, industries are experimenting with virtual receptionists as a core part of their cloud contact center software. But they’re running into a problem that can bring those efforts to a screeching halt: resistance from consumers. Fortunately, a new study shows an important step to overcoming it.
A “common thread” behind much of the resistance involves perception, the study notes. No matter how human-like your agentic voice and other AI technologies may be in their design, people feel that it’s just a veneer. That makes sense, of course, since one could argue that it technically is a veneer. But the researchers test out a solution to help assuage the feelings associated with that awareness.
Building ‘deep humanization’
Many current technologies are achieving “surface humanization,” or “surface anthropomorphization,” the study says. They offer “perceptual cues, such as the presence of a human face, voice, or conversational style. These cues can make an AI appear human-like, but they do not alter consumers’ inferences about the system’s underlying essence of being machine-like.”
The way around this is not to try to fool customers, nor to guide them to forget that they’re talking with AI. In fact, the opposite tactic can work: opening up to customers about how the technology works. Specifically, it can help to inform them that the AI tool was trained on the real experiences of large numbers of people.
This kind of “human training data (HTD) explanation” can trigger “perceived essence transfer,” authors Stephanie Kwari Dharmaputri, Greg Nyilasy, Anish Nagpal and Jing Lei write. With this method, “consumers may build the perception that an AI system has internalized core human qualities, such as thought, emotion, or judgment, through its training process. This form of humanization reflects an attribution of humanness that is internalized, grounded in the belief that human essence has transferred into the technology.”
In my work at Nextiva, I’ve seen what a difference this can make. Our technology powers, and learns from, billions of interactions every year. When some customers are told that, it helps them feel more comfortable. It indicates that even though the technology they’re interacting with isn’t human, it understands people -- including, most likely, people whose concerns, interests, and needs are similar to their own.
The High-Trust Backfire
But another finding in this study is just as important: Not everyone finds this information helpful. When consumers already have “high pre-existing trust in AI technologies,” being given information about the inner workings of the tool can backfire.
“High-trust consumers typically see little need for further justification and are inclined to accept AI output without additional explanations,” the study says. “In this context, HTD explanation may disrupt rather than reinforce confidence.”
Here’s the lesson: Organizations should be ready to offer up information to customers about how an AI tool is working. But they should only do so for some customers. How can an organization know when to do so, and when not to? That's where the power of unified customer experience management (UCXM) comes in.
Unification Breeds Answers
In addition to collecting data from many different people, today’s communication platforms need a complete picture of each individual. They need to know how the customer communicates across every channel, what their journey has been like, how they respond to prompts, and much more.
An agentic voice, or any other AI tool for customer interactions, can use that information to predict the right course of action. It serves as crucial context, showing how lessons learned from other people might, or might not, apply to an individual.
That's why a UCXM platform is the solution. At Nextiva, we’ve seen what happens when we empower XBert (our AI receptionist) with the power of unified data. It can tailor the customer experience to each individual based on their needs.
For agentic voice technology to live up to its potential, organizations need to guide consumers to a point of maximum comfort. That can’t happen in a world of data silos that bring progress to a standstill. With UCXM powering an AI receptionist, a business can proceed into a stronger future, full speed ahead.