#281 What Is AI Audio Separation with AudioShake CEO Jessica Powell

Jessica Powell, CEO of AudioShake, joins SlatorPod to talk about how AI-powered audio separation is making audio more usable for both human and machine workflows, and enabling new use cases across localization, broadcasting, and media production. 

Jessica emphasizes that early traction came from the music industry, particularly in areas like sync licensing and remixing. However, the company’s expansion into film and television happened organically as new use cases emerged.

The CEO explains that AudioShake’s core technology uses source separation to break complex audio into individual components such as dialogue, music, and sound effects. She describes how this allows users to gain precise control over audio for tasks like editing, transcription, and multilingual dubbing.

In localization, Jessica highlights how separating dialogue from music-and-effects (M&E) tracks enables both traditional dubbing and AI-assisted workflows, particularly for legacy content where original stems are unavailable.

Beyond localization, Jessica underscores the importance of clean audio inputs for speech recognition systems. In noisy environments like sports broadcasts or unscripted content, separating dialogue before transcription significantly improves accuracy.

Jessica also reflects on the broader AI landscape, noting that the rise of generative AI has increased awareness of audio as a critical modality. However, she distinguishes AudioShake’s work as non-generative, focused on extracting structure rather than creating new content.

The CEO discusses the current funding environment in the Bay Area and how the investor narrative has evolved leading up to AudioShake’s late 2025 Series A.

Looking ahead, Jessica points to real-time processing and copyright-compliant audio editing as key areas of innovation, as the company continues to expand its role in media and AI ecosystems.