Revolut has rolled out ElevenLabs AI voice agents as its first line of phone support across the UK and Europe, covering more than 4 million customers. The deployment went live this week, according to an announcement on the ElevenLabs blog.
The pitch: premium support without premium costs
The fintech giant, which now serves over 70 million customers globally, wanted to offer phone support at scale without blowing up its cost base. Internal prototypes proved the concept worked, but productionizing the full stack (speech-to-text, LLM orchestration, text-to-speech, real-time turn-taking) would have required platform-level engineering. So they went shopping.
Revolut tested multiple vendors head-to-head. According to Ines Azevedo, the company's Tech Product Owner for AI, some lacked flexibility, others missed reliability targets. ElevenLabs won on voice quality and latency.
The agents handle three jobs: answering FAQs with account-specific context, using live customer data for precise responses, and managing disputes and chargebacks end-to-end. Language detection switches automatically mid-call.
What the numbers actually mean
The headline stat is an 8x reduction in time-to-resolution. That comparison includes queue time for human agents, which does most of the heavy lifting. When you're comparing "instant pickup" to "wait 15 minutes then talk for 5," the math gets dramatic fast.
The 99.7% call completion rate is more interesting. That's 3 failed calls per 1,000, which for an AI system handling financial services queries is reasonably tight. Whether "successful completion" means "customer got what they needed" or "call didn't crash" isn't specified.
31 languages available, though Revolut operates in nearly 40 markets. Coverage gaps presumably get routed to human agents.
Why this matters for ElevenLabs
The startup, now valued at $6.6 billion after a tender offer in September, has been pushing hard into enterprise conversational AI. The Revolut deployment is a marquee case study: a household-name fintech with millions of users and serious compliance requirements.
ElevenLabs hit $200 million ARR by August 2025, according to company statements, with enterprise revenue up over 200% year-over-year. More than 2 million conversational agents have been built on their platform. Revolut represents exactly the kind of deal that justifies a 22x forward revenue multiple.
Julia Ponomareva, Revolut's Director of CX and AI Products, credited the partnership with matching the complexity and speed their markets require. The deployment apparently involved ElevenLabs' Forward Deployed Engineers working directly with Revolut's team.
What Revolut isn't saying
The announcement doesn't mention what percentage of total support volume these agents handle. Covering 4 million customers in the UK and Europe leaves 65+ million others presumably still routed through existing channels.
Chargebacks and disputes are high-stakes interactions. The claim that AI handles these "end to end" deserves scrutiny. Most banks require human review for dispute resolutions above certain thresholds. Revolut's approach here would be worth watching.
No word on customer satisfaction scores, complaint rates, or how the AI performs on edge cases. The 99.7% completion rate is a system metric, not a user sentiment metric.
What's next
Revolut says it's exploring additional applications across the consumer financial journey. ElevenLabs is angling to expand the deployment globally.
The broader implication: if a $75 billion neobank can route meaningful support volume through AI voice agents and hit acceptable performance thresholds, the business case for competitors just got harder to ignore. The race to automate call centers is no longer theoretical.
