The call center: a ubiquitous, often frustrating, but undeniably essential component of modern business. For decades, it's been the domain of human agents, outsourced operations, and endless hold music.
But what if AI could not just augment, but replace the traditional call center model? That's the audacious vision of Retell AI, a startup led by co-founder and CEO Bing Wu.
Fresh off stealth, Retell AI wasn't just building another chatbot. They're developing a full-stack "AI BPO" – an AI-powered business process outsourcing solution poised to disrupt the $350 billion global contact center market.
The story of Retell AI is a classic tale of entrepreneurial agility. As Bing recounts, the company's initial focus was very far from AI-driven customer service. "We were actually initially building a live streaming dubbing software," Bing explains. "We saw a need for foreign live streamers to be able to reach the English audience."
The concept was promising, but the execution hit a wall: a crippling seven-second delay.
"That product had seven seconds of latency," Bing admits. "Basically, you speak one language on the one side. It comes out in English, but we didn't get much traction."
It was a dead end, but also a catalyst. Some user feedback from their online community sparked a pivotal shift. "Along the way, I think we got feature requests and feedback saying that why can't this be generated by AI?"
That question became the company's new mission. Retell AI pivoted to tackle the challenge of creating truly conversational AI agents. The initial focus was on providing an API to enable developers to build their own AI-powered voice applications.
The distinguishing feature for Retell’s technology lies in its ability to create realistic, human-like conversations. A key breakthrough was their proprietary "turn-taking" model, which enables the AI to seamlessly transition between speaking and listening.
"We built our own turn-taking model that can detect when the end of thought is," Bing explains. "And start generating a response whenever we think the end of thought is there. And that dramatically decreases latency and has good control of the conversation."
This very subtle, yet crucial innovation is the key to creating a positive user experience. As Bing points out, the methodologies adopted in traditional approaches for turn-taking often depend on vocal activity detection and temporary silences, which leads to awkward and unnatural conversations. Retell AI works in contrast to these methods: Its model uses semantic and audio cues, mimicking the manner in which people normally talk to one another.
Initially, Retell wanted to supply a voice AI API for developers to build their own applications. However, the company soon recognized an even greater opportunity: to create a fully AI-enabled BPO solution.
"In this way, I would say what we are currently producing is more directly a mid-tier enterprise company growth platform with automated phone calls so companies can quickly achieve gains in productivity." Bing says. Retell focuses on businesses with between 100 and 500 human agents, and claims that by deploying its technology as much as 50-60% of the workforce can be laid off.
Retell’s vision is more than just automating conversations on the horizon. The company is working to develop a full solution that meets needs unique to contact centers, including features like warm transfers, CRM integration, and performance analytics reports.
The key to Retell's success will be how well it's able to link into existing business structures As Bing says, this is not a simple feat. "We definitely need a fair amount of integration. I think the major integration is CRM."
Creation is underway for native integrations with popular CRM platforms like Salesforce and HubSpot. But Bing points out that more customized linkage may be required to mesh with back-office applications and other in-house systems.
Over the long term, Retell’s aim is to produce a system that is both sophisticated and user-friendly. "We aim to be able to provide an end-to-end one-stop contact centre solution," Bing says. "So it's like AI BPO."
In contrast to this long-term vision, Bing is downright realistic: "So far as AI and technology development goes—very short term. Regardless of how difficult a problem or how many layers of problems there might be involved."
He also stresses the need for refocusing on reliable automation of more basic tasks. "We must make sure that even on L1 and L2 tickets, we are giving solutions reliably, traversing through LLM hallucinations while still making sure that the tone is always right... All of those will already be very low hanging fruits and can bring a lot of big impact."
But Bing argues that AI can already replace a significant proportion of offshore BPOs. "If we improve the reliability of our agents, we could already replace 60, 70, even 80 percent of such offshore BPOs." Yet for this to happen, he says, there needs to be follow through and challenges met not only with "hallucinations" (the AI generates wrong facts), but brand voice must stay consistent.
Retell AI is entering a competitive market, but the company's focus on natural conversation, seamless integration, and a full-stack AI BPO solution could give it a significant edge. Wu states, "I think, our job is to bring those models to contact centers to help them solve, you know, more and more advanced issues and problems more reliably."
While challenges remain, the potential to disrupt the massive call center industry is undeniable. As AI continues to advance, Retell AI is positioning itself to be a leader in the next generation of customer service.
Whether it can truly replace human agents remains to be seen, but one thing is clear: the future of the call center is being rewritten by AI. But for now, it seems the AI-powered BPO will be tackling Tier 1 and 2 requests any time soon.
Revolutionize your call operation with Retell.
.jpg)
.jpg)
%402x%20(5).svg){kind=icon}