Call centers rely on automation. There’s debate about whether this is a good thing, but it’s happening – and perhaps it’s accelerating.
According to research firm TechSci Research, the global contact center AI market could grow to nearly $3 billion in 2028, from $2.4 billion in 2022. Meanwhile, a recent survey found that around half of contact Centers plan to introduce some form of AI next year.
The motivation is obvious: call centers want to reduce costs and expand their operations at the same time.
“Companies with strong call center operations that want to scale quickly without the limitations of human contact center agents are highly receptive to adopting effective AI voice agent solutions,” entrepreneur Evie Wang told TechCrunch. “This approach not only reduces overall costs, but also shortens waiting times.”
Wang is a co-founder of Retell AI, which provides a platform that allows companies to create AI-powered “voice agents” that answer customer calls and perform basic tasks like scheduling appointments. Retell’s agents are based on a combination of large language models (LLMs) that are fine-tuned for customer service use cases and a language model that gives voice to the text generated by the LLMs.
Retell’s customers include some contact center operators, but also small and medium-sized businesses that regularly handle high call volumes, such as telemedicine company Ro. You can create voice agents using the platform’s low-code tools or create a custom LLM ( e.g. upload an open model like Metas Llama 3) to further customize the experience.
“We are investing heavily in the voice conversation experience as we see this as the most critical aspect of the AI voice agent experience,” Wang said. “We see AI voice agents not as mere toys that can be created with a few lines of input, but rather as tools that can provide significant value to companies and replace complex workflows.”
Retell worked quite well in my short tests, at least on the calling side.
I scheduled a call with a Retell bot using the demo form on Retell’s website. The bot walked me through the process of scheduling a hypothetical dentist appointment, asking questions like my preferred date and time, my phone number, etc.
I can’t say the bot’s synthesized voice was the best I’ve ever heard in terms of realism – certainly not on par with Eleven Labs or OpenAI’s text-to-speech API. Wang, in Retell’s defense, said the team focused primarily on reducing latency and dealing with edge cases, such as interruptions that might occur in a conversation.
The latency Is low: In my test, the bot responded to my answers and follow-up questions almost without hesitation. And it stayed true to its script. Try as I might, I couldn’t confuse it or cause it to behave that it shouldn’t. (When I asked the bot for my dental records, it insisted that I speak to the office manager.)
So are platforms like Retell the future of the call center?
Perhaps. For basic tasks like scheduling, automation makes a lot of sense, which is probably why both startups and large tech companies offer solutions that directly compete with Retell’s. (See Parloa, PolyAI, Google Cloud’s Contact Center AI, etc.)
It’s a low-hanging – and seemingly revenue-generating – fruit. Retell claims to have hundreds of customers, all of whom pay per minute of conversation with a voice agent. Retell has raised a total of $4.53 million in capital to date, courtesy of backers like Y Combinator (where the company was founded).
But on more complicated questions, the jury is out, especially given the tendency of LLMs to make up facts and go off track even when safeguards are in place.
As Retell’s ambitions grow, I’m excited to see how the company overcomes the many well-known technical challenges in this area. At least Wang seems convinced of Retell’s approach.
“With the advent of LLMs and recent breakthroughs in speech synthesis, conversational AI is becoming good enough to create really exciting use cases,” Wang said. “For example, with sub-second latency and the ability to interrupt the AI, we observed users speaking in more detailed sentences and conversing as they would with another person. We’re trying to make it easier for developers to build, test, deploy and monitor AI voice agents, ultimately helping them reach production readiness.”