Ai-centric enterprise communications company Sendbird this week added to its arsenal with a voice-automation tool. Known as Senbird Voice AI, it’s meant to handle inbound customer service and offload the manual rigor that goes into it. It boasts features like response speed and natural voice interaction.
Taking those one at a time, speed is a key factor in automated voice systems. Customers have little patience for automated voice-response systems, which are widely disliked and bypassed. So speed is the name of the game in processing questions and providing meaningful answers or call routing.
This comes as many conversational AI systems fall short on speed. Sendbird specifically cites Forrester’s State of AI Agents report, which asserts that even as AI adoption increases in customer service, most providers today underperform. This boils down to subpar speed and consistency in helping customers.
As for natural voice interaction, Sendbird claims it can cut through the “messiness” of real conversations. By that, it means that Voice AI is built for the unpredictability of authentic human communication. So it focuses on processing nuances in human speech like accents, inflections, and interruptions.
Specifically, the full list of features offered in Voice AI is…
- An AI inbound voice agent with real-time analytics and transcripts;
- Text-to-speech and speech recognition in 29 languages;
- Accent, language, and tone customization;
- Development environment testing and transparency into AI decision making; and, Smooth AI-to-live agent transfers, when escalations from an AI agent to a human agent are necessary.
- In addition, every deployment is tested against realistic scenarios to ensure it understands intent, adapts on the fly, and resolves issues as expected with Sendbird Trust OS.
Balancing Act
Zeroing in on that last part, Trust OS is Sendbird’s recently-launched accountability system that makes AI agents more trustworthy. As we wrote at the time of its launch, this is all about counterbalancing AI’s trust issues. Though ample buzz surrounds AI, it’s too early and unproven to trust it with sensitive data.
Trust OS addresses those concerns with functions that breed accountability, oversight, and responsible agent behavior. For example, it offers a dashboard that lets operators look over every AI agent decision and interaction to ensure things are going smoothly. Vendasta recently rolled out something similar.
We bring up Trust OS because these announcements are all interrelated. Because Voice AI is inherently customer-facing, it’s one of those areas of AI where enterprises are still wary to integrate it. Trust OS provides that extra security blanket or system of guardrails to instill confidence in AI deployments.
Back to Voice AI, it’s also about a balancing act. AI offers enterprises opportunities to automate and streamline operations, which can generate cost efficiencies when done right. But too much automation in customer-facing ways can be detrimental. So platforms like Voice AI have to accommodate.
Specifically, it’s all about automating rote parts of customer interactions that can and should be automated, while preserving human touch when needed. That defines the past few decades of customer service automation, but it’s in hyperdrive in the AI age. We’ll see if Sendbird strikes the right balance.
