Customer support · Production

Revolut deploys ElevenLabs Agents to cut time-to-resolution more than 8x in customer support

The problem

As Revolut expanded globally across nearly 40 markets, the team needed to offer premium voice support at scale without sacrificing cost discipline, compliance, or control—but productionizing the full AI stack in-house proved to be a platform-scale lift.

First attempt

Revolut's internal AI build validated the concept but could not be productionized at scale due to the complexity of speech-to-text, LLMs, TTS, real-time turn-taking, PCI compliance, and zero-retention controls. Other evaluated vendors lacked flexibility, missed reliability targets, or could not match the required voice quality.

Workflow diagram · grounded in source
1
Customer initiates voice support call
Trigger
ElevenLabs Agents serve as the first line of voice support for customers in the UK and Europe.
▾ source quote
“Revolut deployed ElevenLabs Agents as its first line of voice support for customers in the UK and Europe”
2
Real-time language detection and switching
Ai action
The agent detects and switches languages in real time during live calls.
▾ source quote
“The ElevenLabs agent handles live calls, detects and switches languages in real time”
3
Secure connection to proprietary systems
Integration
The agent connects securely to proprietary systems to handle its three core support jobs.
▾ source quote
“connects securely to proprietary systems to handle three core jobs”
4
Account-specific FAQ and service answering
Ai action
The agent answers FAQs about accounts and Revolut services using live customer data to give precise, not generic, responses.
▾ source quote
“answer FAQs about accounts and Revolut services with account-specific context, use live customer data to give precise - not generic - responses”
5
End-to-end dispute and chargeback management
Ai action
Wherever possible, the agent manages disputes and chargebacks end to end.
▾ source quote
“wherever possible, manage disputes and chargebacks end to end”
Reported outcome

ElevenLabs Agents cut time-to-resolution by more than 8 times, scaled coverage to more than 4 million customers in more than 30 languages, and achieved a call success rate of more than 99.7% with tickets resolved in less than 5 minutes.

Reported metrics
Time-to-resolution improvementmore than 8 times
Customers covered by voice supportmore than 4 million customers
Languages supportedmore than 30 languages
Call success rateMore than 99.7%
Show all 6 reported metrics
time-to-resolution improvementmore than 8 times
customers covered by voice supportmore than 4 million customers
languages supportedmore than 30 languages
call success rateMore than 99.7%
ticket resolution timeless than 5 minutes
previous agent resolution time comparisonprevious agents took 8x the time or more
Reported stack
ElevenLabs Agentsspeech-to-textlarge language modelstext-to-speech models
◆ Does this fit your context?

Compare to your context

Tell us your scale, team, and constraints. We'll show what changes at your size, what fails at your scale, and whether this case is a fit, needs adaptation, or won't scale to you. Free demo, no signup.

Compare to your context →
~30 seconds · free
Source
https://elevenlabs.io/blog/revolut
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

ElevenLabs Agents cut time-to-resolution by more than 8 times, scaled coverage to more than 4 million customers in more than 30 languages, and achieved a call success rate of more than 99.7% with tickets resolved in l…

What tools did this team use?

ElevenLabs Agents, speech-to-text, large language models, text-to-speech models.

What results were reported?

Time-to-resolution improvement: more than 8 times; Customers covered by voice support: more than 4 million customers; Languages supported: more than 30 languages; Call success rate: More than 99.7% (source-reported, not independently verified).

What failed first in this deployment?

Revolut's internal AI build validated the concept but could not be productionized at scale due to the complexity of speech-to-text, LLMs, TTS, real-time turn-taking, PCI compliance, and zero-retention controls.

How is this customer support AI workflow structured?

Customer initiates voice support call → Real-time language detection and switching → Secure connection to proprietary systems → Account-specific FAQ and service answering → End-to-end dispute and chargeback management.

Related customer support cases

WHAT TO DO WITH THIS

Now compare it to your context

This case is one data point. Whether its pattern fits you depends on your volumes, your stack, and your exception load — that comparison is the step no case study can do for you.

Browse the Customer support cases →