A new podcast-generation capability has surfaced inside Microsoft Copilot, allowing the assistant to turn almost any topic or user-supplied materials into a brief, conversational show. The tool scripts a back-and-forth between two synthetic hosts—an approach that closely echoes the “Audio Overview” format Google debuted last year—so listeners receive a dynamic summary rather than a single-voice narration.
Microsoft outlined the feature during its 50th-anniversary Copilot event on 4 April 2025 and in a follow-up blog post, explaining that it can curate personalized programs covering everything from vacation planning to academic papers.
In use, a person chooses a topic or pastes source content, Copilot analyzes the material, drafts a script, then renders the dialogue with its neural text-to-speech stack. Playback stays interactive; pausing to ask a question causes Copilot to adapt the conversation in real time, streamlining multitasking and promising a better user experience.
AI-generated podcasts are not entirely new—Google’s NotebookLM and Spotify’s year-end stories pioneered similar experiences—but adding them directly to Copilot lets Microsoft compete more effectively while enhancing multimodal output for hundreds of millions of Windows and Bing users. Early hands-on reports note audio that still trails NotebookLM, yet the move represents a significant enhancement that could improve user engagement through richer interactions.
Microsoft says Copilot Podcasts will roll out “over the coming weeks,” with enterprise, mobile, and API access slated for later this year, ensuring broad availability once testing concludes.
