Could you tell the difference between a human customer assistant and a tool powered by artificial intelligence (AI)? Vapi , which is today announcing it has raised $50 million of Series B funding, says its voice technology is so good that many users won’t notice they’re talking to such a tool.

Companies are increasingly pinning their hopes on such technology as they battle to improve customer satisfaction – as measured by metrics such as CSAT scores – that have continued to fall despite significant investment in a range of tools in recent years. Many customers report frustration with automated call handling services, call center delays and unwieldy chatbots; Vapi says its voice agents can overcome such issues.

“Voice represents an opportunity to bring your costs down but more importantly, we can help you consistently function at the level of your very best human,” says Jordan Dearsley, CEO and co-founder of the company. “The real unlock is building agents for your customers that feel human.”

Vapi works with its customers to do exactly that: to build and deploy voice agents that can handle a broad range of customer inquiries, rapidly increasing an organisation’s capacity to deal with customers but without compromising satisfaction. Once implemented, the AI tools can autonomously manage most queries, but they also improve over time as the underlying models learn from each call handled.

I first profiled the San Francisco-based start-up in December 2024, just as it was putting the finishing touches on a $20 million Series A funding round. Since then, Vapi’s revenues have grown 10-fold, thanks to deals to supply its voice agents to customers including Amazon, New York Life, Intuit and Service Titan. The company has just announced its voice agents have now handled 1 billion calls.

Still, it has taken time to convince people that voice agents really can deliver what they promise, concedes Dearsley. “I think there is a general scepticism about AI projects because people have been burned by disappointing outcomes,” he says. “But businesses are rational – when they see the returns on investment that voice agents can deliver, they’re keen to go further.”

That’s been the experience of Amazon Ring, says Jason Mitura, the company’s vice president of software development. “100% of our inbound volume now runs through Vapi,” he says. “Most importantly, we’ve maintained our high bar of support for our customers and CSAT scores have improved.”

It’s certainly a fast-growing industry. Research from MarketsandMarkets estimates the AI voice sector was worth $4.2 billion globally last year but expects that figure to rise to $20.7 billion by 2031, representing annualized growth of almost 31%. Competition is becoming more intense, with Vapi’s rivals, including firms such as Bland, Millis and Retell, getting attention too.

Hence Vapi’s need to accelerate at an even faster pace. Dearsely is now focused on improving the experience in three areas. First, he points out, achieving reliability at scale is tough for voice agents, requiring significant investment to ensure that quality doesn’t slip as agents handle more calls. The second priority is compliance – particularly in industries such as health and financial services, where companies can’t afford a misstep because a voice agent functions poorly. The final area getting attention is performance, ensuring that the AI models underpinning voice agents learn more quickly to drive up returns on investment.

Vapi will therefore use the proceeds of today’s Series B round to invest in further engineering expertise, as well as in go-to-market resources. The round is led by Peak XV with participation by the Microsoft venture fund M12, Kleiner Perkins, Bessemer Venture Partners and a number of the company’s earlier investors. The round brings the total amount of funding raised by Vapi to $72 million.

Arnav Sahu, partner at Peak XV, believes that as the voice market expands, Vapi can emulate the success of technology leaders in other areas of enterprise software. “In 10 years, it’s likely most calls will not have a human behind the phone,” Sahu argues. “With its bottom-up, product-led growth approach, we believe Vapi is the next Zapier and N8N for voice AI workflows.”