Spoken Automates Multi-Cast Audiobook Production with AI

Spoken, founded in 2024 by medical technologist and entrepreneur Phil Marshall, has developed an AI-powered audiobook platform that analyzes a manuscript's language, style, narrative structure, and characters to automatically assign distinct voices for narration and dialogue. The platform supports single-narrator, dual-narrator, and multi-cast productions. Authors can choose from over 100 paid professional voice actors, clone their own voices, or use custom-generated voices. To prepare the text, Spoken utilizes an agentic AI workflow where the computational power used to analyze and prepare the manuscript is about five times the amount of AI used in the actual voice narration.

A recent study by Edison Research highlighted the platform's performance by comparing Spoken's multi-cast production against a professionally produced, single-narrator human version. Before hearing the samples, only 31 percent of participants expressed interest in AI-narrated audiobooks. However, after hearing excerpts of Spoken Multi-Cast without knowing it was AI, 65 percent of listeners said they would listen to an entire audiobook with that narration. Furthermore, only 39 percent of those who heard the Spoken Multi-Cast suspected it was generated by AI, which was highly comparable to the 35 percent of listeners who mistakenly suspected the human version was AI. Once participants were informed the narration was AI-generated, overall willingness to listen to AI audiobooks rose to 43 percent.

For authors and publishers, this technology dramatically lowers the cost of producing high-quality audiobooks, which traditionally require expensive recording studios and voice actors. Spoken offers an application programming interface for publishers managing large catalogs. Authors retain full ownership of their work, masters, and distribution rights. To maintain creative control, creators can use a Speak It feature to demonstrate specific inflections or pronunciations into a microphone, which the AI voice then replicates. Spoken also maintains an ethical AI model, ensuring written works are not used to train its systems and participating voice actors, including those sourced from ElevenLabs, are compensated for their work.