We explored nearly 10 new types of audio formats that could be generated at scale using AI. From this bank of initial proposals, two were refined and turned into usable workflows for O’Reilly Media’s production team.
HONORING AN ARCHIVE OF EXTENSIVE LEARNING MATERIAL
As a leading voice in digital technology literacy, O’Reilly Media was uniquely positioned to create AI functionality in an educational context. Its expansive pool of resources spanning books, tutorials, and training programs presented an opportunity to innovate new AI-driven learning experiences that centered their learners’ personal journey while staying true to their roots and values.
LESSON PLAN FOR THE LONG HAUL
Sustaining users for a long-form education service like O’Reilly Media requires constant experimentation and alternate consumption patterns. We explored new types of audio-led educational materials that can be generated at scale with AI on a fast-tracked timeline without the need for custom development.

Book Summaries
Introducing chapters in a book through different voices and dynamically-paced content structure opened up a world of possibilities. This new content type introduces a book and its chapters in a dynamic format1 with different voices.
We hear an introduction to the book, a summary of each chapter2, and a summary of the conclusions. Actual snippets of the book are read aloud in a different voice to create a more dynamic format. Runtime is usually around 10–15 minutes.
Topic 101s
A podcast show that features a basic introduction to a new topic or sub-topic from the learning platform in every episode. We sourced introductory chapters from books3 available on O’Reilly Media’s platform.
These chapters were summarized into single introductions4 and featured background music to support the text-to-speech narration. Each podcast episode saw a runtime between 15-45 minutes.