A coordinated audio-visual deepfake combines face swap and voice clone synchronously during a live video call, producing a complete real-time impersonation. The 2024 Arup case, where an employee transferred USD $25 million after a multi-participant video call deepfake, is the canonical example.

• Face swap runs in real time on the attacker’s webcam feed
• Voice clone or voice conversion is applied to the audio stream
• The two are synchronized so lip movement, prosody, and timing match throughout the call

• Live wire transfer authorization on Zoom or Teams calls
• Synthetic identity verification during remote KYC
• Fake board or executive meetings used to extract decisions or documents
• Compromised vendor calls where the spoofed counterparty appears legitimate
• Insider threat scenarios where the attacker has impersonated a colleague