From Text to Discussion: Gemini's New Feature Turns Reports into Podcasts in Seconds

Google has merged two key technologies — Deep Research and Audio Overviews—to create dynamic audio podcasts. Now, users can turn complex analytical reports into engaging conversations between virtual hosts, opening up new possibilities for learning and working with information.

The previously launched Deep Research feature allows Gemini to analyze massive amounts of data from the internet and generate detailed reports on a given topic. Starting today, this tool now includes the Generate Audio Overview option, which transforms dry facts into an interactive podcast. Once the research is complete, the user selects the appropriate button, and Gemini generates a dialogue between two AI "hosts" discussing key findings in a lively conversation format.

The voiced summary technology first debuted in September 2023 in the NotebookLM app, designed for document analysis. Over the past year and a half, Google has significantly improved the feature: users can now influence the script of the dialogue, ask virtual speakers questions, and even change the style of the content delivery. In March 2025, the service became available in Gemini for everyone — both free users and subscribers.

While the system previously worked only with short texts and presentations, integration with Deep Research now allows it to process full-scale research — ranging from climate reports to market analyses. Experts note that the transition from monotonous textual summaries to interactive podcasts solves the issue of "information fatigue." The two AI conversationalists simulate a discussion, highlighting contradictions, supporting arguments with examples, and even adding storytelling elements. According to Google representatives, this format improves data retention by 40% compared to traditional methods.

By the end of 2025, Google plans to add the ability to choose host voices, adjust episode lengths, and insert user-generated questions in real-time. The new era of personalized audio content, where every request becomes a unique show, has already begun.