Why voice is still the most human channel: A conversation with Sendbird CEO John Kim
When conversations matter most, people still pick up the phone. Despite the rise of messaging apps, live chat, and email, voice remains the channel people trust when the stakes are high. It is immediate, instinctive, and deeply human.
We sat down with John Kim, our CEO and co-founder, to talk more about the enduring relevance that makes voice one of the most important frontiers in customer experience. “Voice is the most universal user interface,” he says. “People talk faster than they type. And for many, it’s simply more natural. Whether it’s the elderly, those with limited literacy, or people who just don’t want to download another app, voice opens the door to conversations that text alone can’t reach.”
People still pick up the phone when it matters most
The power of voice lies in both its speed and accessibility. Billions of people can speak long before they can write, and in many regions typing remains a barrier to digital communication. Voice bypasses those hurdles and creates an inclusive way for more people to engage. John also points out that spoken communication is more forgiving than text. Typos and grammar errors interrupt a message when reading, but when listening our brains fill in the gaps. That flexibility makes voice especially valuable for older generations and in multilingual contexts where text feels rigid.

It’s why voice continues to hold its ground even as digital channels multiply. When something breaks, when customers are frustrated, or when the problem is too important to risk miscommunication, people still prefer to talk. The ability of voice to carry tone, emotion, and nuance ensures that it remains the most trusted channel and the one best equipped to resolve issues quickly.
The challenges of making machines sound human
Yet voice has been one of the hardest channels for AI to master. Early attempts at automation left customers with more frustration than answers: robotic voices that stumbled over natural language, awkward silences that made the interaction feel broken, and limited multilingual support that excluded global audiences. Customers often ended those calls exasperated, demanding a human agent or simply hanging up.

John likens this challenge to the “uncanny valley” of computer graphics. “When an AI voice doesn’t sound natural, it feels creepy. People don’t just notice—they reject it.” Latency adds to the problem. A few seconds of silence may seem minor in computing terms, but in a conversation it feels unbearable, undermining trust in the entire experience. These flaws explain why earlier generations of voice automation never achieved lasting traction.
Designing voice AI for real conversations, not scripts
Sendbird has reimagined how voice AI agents should work by focusing on the details that make conversations feel authentically human. Its agents are designed to respond instantly, weaving in natural conversational cues so dialogue flows the way it would with a real person. They handle interruptions, adapt when customers change direction mid-sentence, and filter through background noise without losing track of intent. Rather than confining interactions to rigid scripts, the system is built for the chaos of real conversations.
Equally important is multilingual performance. With customers across every region, Sendbird has invested in making its voice AI agents sound natural in dozens of languages and dialects, not just English. And because voice is integrated into Sendbird’s omnichannel platform, it never exists in isolation. If a call drops, the interaction can continue seamlessly over SMS, chat, or email with the full context carried forward. “Once you’ve experienced this kind of continuity, you can’t go back,” John explains. “No one wants to repeat themselves every time they switch channels.”

The next generation of voice experiences
Looking ahead, John believes the next leap in voice AI agents lies at the intersection of brand identity and personalization. A conversation with Disney should not sound like a conversation with IBM, and an AI agent handling a frustrated customer should not respond the same way it does with someone making a casual inquiry. “Each brand has a distinct voice, and each customer situation calls for a different response,” John says. “Voice AI agents should be able to capture both.”
He also sees voice becoming an important bridge between AI agents themselves. Some machine-to-machine communication will happen through efficient data protocols, but voice offers a universal interface that keeps those exchanges accessible to humans. In the long run, AI agents may evolve their own optimized languages, but until then, voice remains the most intuitive way for people to connect with technology.
Voice still matters and always will
Voice is not disappearing. If anything, it is becoming more central to how people interact with businesses, technology, and AI itself. Sendbird is building voice AI agents that don’t just answer calls but listens, adapts, and carries a brand’s identity into every conversation. For John Kim, that is what makes voice not just another feature, but a defining part of the future of customer experience.
Ready to hear it for yourself? Request a demo and experience the future of voice in action.