By Yaniv Masjedi, Chief Marketing Officer at Nextiva
When a customer calls your company, would they want to hear their own voice on the other end of the line, speaking on behalf of the business? Probably not. Voices cloned by AI may have their uses, like helping people express themselves across multiple languages. But it’s unlikely that many customers would want to experience a vocal doppelganger effect.
However, customers may prefer voices that are like their own. In fact, according to new research, they do.
That’s the kind of finding businesses need to pay attention to as they adopt AI receptionists as part of their cloud contact software. Voice agents are now the second most used AI channel in customer service, according to COPC.
In a hyper-competitive era, organizations are seeking out every advantage they can get to improve the customer experience. A customizable voice assistant can go a long way in managing calls more effectively.
“As voice agents become embedded in products as well as service interactions, firms increasingly seek personalization strategies that strengthen consumers’ acceptance and engagement,” the new study says. “This research examines whether voice agents that sound more similar to users’ own voices elicit more favorable behavioral responses.”
The upshot: Yes, they do. The researchers found that “higher voice similarity increases purchase intention, use intention, and willingness to disclose personal information.” The similarity enhances trust as well as “self-congruency,” a term referring to an alignment between a person’s self-image and the image they associate with a product or brand.
This study does not provide a one-size-fits-all solution. But it highlights something crucial for business leaders to understand as they enhance their communication platforms with the latest technology. More types of personalization are available today than ever before. AI-powered call centers can fine-tune everything from voices to intonations, inflections and more.
To develop the right voice for each individual, a business needs to understand the person’s unique preferences. For example, some customers may prefer voices that evoke subconscious associations with other people in their lives, or voices that have characteristics they associate with trust. The key is to remove the guesswork, and take cues from the individual.
Fortunately, that can be done automatically. One way is through unified customer experience management (UCXM), which involves pulling together every piece of information a business has about a customer, across all channels. In every interaction, people indicate their preferences. And when they communicate by voice, their vocal characteristics can be analyzed.
To be clear, none of this can be done in a way that violates a customer’s privacy. The new study notes that, “Compared with typing, disclosure via speaking tends to carry higher privacy expectations because people feel the information is more personally owned by them. Therefore, if firms leverage self-similar voices to increase consumers’ willingness to disclose, they also bear heightened responsibility to ensure data handling matches these expectations.”
To whatever extent customers allow their data to be used to improve the customer experience, it’s up to the company to maximize the power of that information. In numerous ways, customers show you what works best for them. Their cues come not only from the language they use, but also in vocal mannerisms and other cues. Sentiment analysis tools track these, and follow how customers respond to slight differences in the tone, style, or language the organization is using.
This information can then be combined with data on how other customers with similar characteristics respond to different agentic voices. All this information becomes like a treasure map. Step by step, it leads the organization to the agentic voice that's right for that customer, helping build a stronger relationship with them for the long term.
Achieving all this requires maximum flexibility. AI receptionists should be customizable with multiple AI voice options, adjustable response behaviors, and more, all with granular control. And since sales and marketing teams shouldn’t have to become experts in all this, the AI-powered platform should offer guidance at each stage.
There are times when voice cloning is useful -- but, at least for now, it’s not about cloning customers’ voices. Instead, some people lend their vocal skills to businesses specifically for cloning, so their real conversational style is used by voice agents like ours. When used, these kinds of technologies require staunch safety guardrails.
As people get used to AI, their preferences are likely to change. So it’s important to remain malleable in developing agentic voices. But there's no need to wait to take advantage of what AI receptionists offer today. By combining the power of AI with the need for hyper-personalization, brands can communicate with customers in unprecedented ways, delivering the best CX of their lives.