Zepo uses ElevenLabs to power realistic vishing simulations at scale

Zepo is a Spain-based cybersecurity platform that runs human risk simulation campaigns to reduce organizational exposure to social engineering attacks. Their simulations span multiple channels and vectors, including audio scenarios for phishing and deepfake attacks. To replicate trusted figures convincingly - such as a CFO requesting a wire transfer - Zepo needed natural-sounding, multilingual voice at scale. They integrated Text to Speech through the ElevenAPI platform to meet this need.

Up and running within days

Clear documentation and a low-friction developer experience meant Zepo was generating dynamic audio within days of their first API call, with minimal disruption to existing workflows.

Voice realism as a training tool

Vishing - voice-based phishing - works by impersonating a trusted figure, such as a company executive, to pressure employees into transferring funds or disclosing credentials. The closer the simulation is to a real attack, the more effectively it drives behavioral change.

Using ElevenLabs, Zepo generates audio simulations that closely replicate the vocal patterns of trusted figures across multiple languages, with campaigns designed to reach 15,000–20,000 employees in a single day across Spain and Latin America.

"At Zepo, we believe people are the first line of defense in cybersecurity. ElevenLabs gave us the ability to make our vishing simulations sound indistinguishable from a real call — and that's when behavior actually changes." — Enrique Holgado de Frutos, PhD, COO and Co-founder, Zepo

Early results and projected impact

After four months, client demand has driven significant expansion in campaign volume. ElevenLabs is projected to help Zepo boost engagement and upsell rates in simulated attack training modules by over 50% in the first 9–12 months, with the strongest impact in high-risk sectors such as finance and healthcare.