ElevenLabs' AI SDR Agent Qualified 3,000 Leads in 32 Languages

ElevenLabs' inbound sales development representative (SDR) agent has handled more than 3,000 customer leads this year, qualifying over 90% of them in 32 languages without a human in the loop. That number, disclosed by field chief technology officer Alex Holt, is the clearest signal yet of how far the voice AI startup has pushed agentic AI into revenue-critical work — and Holt says the bigger gains are still ahead.

"You know you always feel like you're on the tip of the iceberg," Holt told Fortune. Today's efficiency wins may feel cutting edge, but "in six months time, you can see even more gains."

What changes for sellers and recruiters?

The SDR agent uses voice AI to gauge how serious a prospective buyer is, then loops in a human rep after screening. In one recent case, a customer closed a sale entirely through two agents: the SDR agent fielded the inbound request, and an account executive agent handled the legal documents and contracts.

"The value of that is you can now automate that whole chain for customers where it's the simple, most repetitive mechanical cases," says Holt. Humans get redirected to complex inbound requests and upselling — a workflow shift that matters most for high-volume, low-complexity deal motion.

The same logic runs through recruiting. Nearly all prospective candidates are now initially screened by an AI recruiter conducting first-round interviews. "That's 30 minutes that the recruiter can spend at the end of the process, convincing you to join," Holt says.

Onboarding has sped up too: a coaching agent lets new sales hires role-play customer interactions before they face real buyers.

How is this organized internally?

ElevenLabs embeds an engineer on every team, and every group stays under 10 people. In recruiting, two forward-deployed engineers sit in a pod with HR colleagues with one objective: cut the time it takes to move candidates through interviews.

Most employees work collaboratively with agents, and a majority have built their own for repetitive tasks. But the business metrics don't change. "It's exactly the same business metrics," says Holt. "This has to be an initiative that the business units themselves are driving and pushing. The agents and tech are a tool that they can leverage to do that." Sales teams are expected to generate more revenue faster; customer service satisfaction scores stay under close watch.

What about governance and mistakes?

Agentic AI comes with a measured caveat: surveys show most businesses remain ill-prepared to rethink workflows and orchestrate multi-agent processes, where one agent autonomously hands work to another. BCG research puts agentic AI's share of AI value at 22% this year, up from 17% in 2025, projecting 39% by 2030.

ElevenLabs runs a human-led monitoring and observation layer over its agents — a direct response to the risk of unintended behavior, which has drawn alarm after high-profile incidents where OpenAI and Anthropic agents went rogue. Mistakes still slip through. One customer service agent erroneously issued a small refund under $30 and failed to escalate. The company built a test case simulating the call and the exact chain of reasoning that led the agent astray, then updated prompts and guardrails.

Holt argues that auditability is a hidden benefit: a human making the same mistake would likely have gone unnoticed. "It's not just automating processes; it allows you to collect more data," he says. "I expect over the next few years... people talking much more about optimizing these processes."

What's next?

Patience is part of the deal. Whether the recruiting agent hires the right people, or service scores improve — "some of these metrics can take a long time," Holt says, and business owners own identifying the leading indicators worth tracking.

The next frontier is multi-agent systems with shared context and more complex handoffs. Holt, who joined ElevenLabs as an engineer in September 2023 after five years at Palantir and was elevated to field CTO in June 2026, is building toward exactly that. Three months after his promotion, the startup doubled its valuation to $22 billion, and CEO Mati Staniszewski has signaled a possible IPO in 2028.