Google has a wildly underrated research assistant called NotebookLM. Ironically, it didn’t exactly take off because of its research capabilities. Instead, the tool went viral due to its Audio Overviews feature, which lets you turn sources into AI-generated podcasts. The podcasts that NotebookLM generates are engaging and, surprisingly, don’t have the robotic voice most AI tools default to. The two hosts sound natural and almost like you’re listening to two humans discussing a document you sent their way.

Though these Audio Overviews have been powered by Gemini since Google added them to NotebookLM in September 2024, the feature was initially restricted to NotebookLM. In March 2025, Google finally made Audio Overviews available directly in Gemini’s interface. Now, the feature is coming to Google Search too.

Google is testing Audio Overviews directly in Search

As announced in a post on Google's The Keyword blog, Google is launching Audio Overviews in Search as a Google Labs experiment. The blog post explains that Search Audio Overviews will use the latest Gemini models to generate quick and conversational audio summaries for certain search queries.

Based on the image that Google shared, it seems like the experiment will display a short Audio Overview on the search results page, right below the People also ask section.

Though it isn't widely available yet, anyone can try it out by opting into the experiment in Labs. You can do so by heading to this Google Labs webpage and hitting the Turn On button. Then, toggle on the switch next to Turn this experiment on or off and hit the Start searching button.

Doing so will essentially turn on all the AI Overviews experiments Google Labs is currently testing, including increased AI Overviews in Google Search. Google mentions that features from these experiments may appear on the search results page in the Google app, a few mobile browsers, and on Chrome desktops. Audio Overviews will only appear on search results pages where Google's systems determine they’ll be helpful.

Since Audio Overviews in Gemini are typically much shorter than the ones you can generate in NotebookLM, I believe it’ll be a similar case in Search. I’m also curious to see if they’ll be as well-liked as NotebookLM’s. I assume these will be generated once a user clicks a certain trigger or prompt, which makes me wonder if a user would willingly wait around for an audio summary to load, especially when it'd take them less time to just skim text.