Google has had its fair share of weird AI experiments. Its dedicated Google Labs website showcases these experiments and lets users take a few of them for a test run. For instance, Google's latest experiment is Little Language Lessons, which are essentially three bite-sized experiences built using the company's beloved AI assistant, Gemini.

The experiments are more or less Google's take on Duolingo and are designed to gamify the process of learning a new language using AI. While a lot of Google Labs experiments slowly fade away, some manage to make it out alive and into the public eye. NotebookLM, Google's AI-powered note-taking app, which was introduced under the code name "Project Tailwind" in 2023, is one of the few that survived.

What made NotebookLM truly stand out from the hundreds of AI note taking tools already out there, and helped it gain traction, was the ability to create AI-generated podcasts. While that's the feature that put it on the map, there are a handful of underrated features quietly built into NotebookLM that I neglected for the longest time, and you might have too.

3 Interactive Mode in Audio Overviews

Why not become the third host of the podcast?

If you've used NotebookLM before, chances are you've tried the Audio Overviews feature at least once. Audio Overviews is simply NotebookLM's fancy term for AI-generated podcasts where two hosts discuss the sources you add to your notebook in an engaging and lively manner (even if your sources are about the most boring topic ever). While these AI-generated podcasts are a great way to quickly overview what's covered in your sources without having to go through each one, they aren't meant to be a comprehensive breakdown. Naturally, there's a chance that NotebookLM's Audio Overview might end up missing that one specific explanation you were hoping to hear.

So, if you're like me and rely heavily on NotebookLM when you're cramming for yet another exam, it may not be the best idea to treat Audio Overviews as your only study method. While you always have the option to add instructions, like asking the hosts to focus on a specific topic by clicking the Customize button, wouldn't it be better if you could just hop into the podcast and ask the hosts your query? NotebookLM's Audio Overview's Interactive Mode lets you do exactly that.

I find doing this much better than the alternative, which is searching for my query online, since there's a chance the answer I get might not be relevant to my coursework. I don't need to worry about that with NotebookLM, though, since the AI is self-contained and remains grounded in your sources, meaning the only knowledge it essentially has is... the sources you provide and your conversation history.

Once you generate your Audio Overview, click the Interactive mode (BETA) button to hop in. Hit the Play icon to get the podcast going, and whenever you'd like to interrupt, simply hit the Join button. When you do, the hosts will call on you by saying something to indicate they're ready to hear you. For instance:

Oh, our listener has something to share, let's hear it.

Woah wait, someone wants to join. Go for it!

Woah, did someone just break the fourth wall? Let's hear it.

Hey there, wanna share your thoughts on this?

Google mentions that since the feature is still experimental, the AI hosts may end up pausing awkwardly right after you pose your question or share a thought. I've used the feature a couple of times while studying, and it's pretty much like having a study buddy on call with you who knows everything and has a knack for cracking weird jokes.

2 Follow-up questions to enhance your understanding

One question is all it takes

As mentioned above, Audio Overviews are great for quickly reviewing and revising the content in your sources, but they aren't ideal if you're studying a topic from scratch. In such situations, I typically ask NotebookLM questions related to my sources in the Chat panel. This saves me the time of reading through my notes, and since NotebookLM is grounded in the sources I’ve added, I can rest easy knowing the answers won’t include any hallucinated information.

Given that I tend to do the above without thoroughly going through my sources, there's no way I can think of every relevant question to ask. Fortunately, NotebookLM helps me out here, too. Once the AI tool answers a query, it suggests a few follow-up questions related to your sources.

While the follow-up questions aren't always directly linked to the query I asked and usually come from other parts of my sources, they still help me cover concepts I may have missed in a hurry. What I've been doing lately is trying to guess the answers to the follow-up questions NotebookLM suggests before clicking on them. This way, I can test what I've picked up so far without needing to open my notes.

1 Discover Sources

Let NotebookLM do some of the heavy lifting

While one of the reasons why NotebookLM is my go-to AI tool is that it doesn't resort to searching the internet when I ask it a query, I’m not against AI resorting to the internet. But it should be for the right reasons. The Discover Sources option in NotebookLM is exactly what I’m talking about. Instead of the AI pulling in random information from sources across the web to answer queries you have about the sources you provide it with, Discover Sources resorts to the web to help you, well, discover sources. All you need to do is open a new or existing notebook and click the Discover button next to the Add source button in the Sources panel.

Simply describe the kind of sources you’d like to add to your notebook, and hit Submit.

The tool then uses Gemini to search through hundreds of web sources (including community forums like Reddit) within seconds and shortlists ten recommended sources it believes are most relevant to your topic. The source’s title, a brief description of how it relates to your query, and a link to the source will be displayed too, making it easy to skim through and decide which ones to add to your notebook. Finally, check the box next to each source you’d like to add, and click Import.

I’ve been playing around with this feature since NotebookLM introduced it, and it’s been a game-changer. I was initially worried that it would cause NotebookLM to start hallucinating information, but I’ve yet to come across that. I find that it’s a great way to expand your notebook with relevant materials. Since you’re still the one deciding which sources get added, you retain full control over what the AI knows and references.

You're missing out if you haven't tried these features

While the tool's more well-known features, like Audio Overviews and Mind Maps, are undoubtedly incredible and are enough reason to try out NotebookLM, and I can't begin to imagine studying without them, each feature above is equally important in my study routine. I just wish they received more attention than they currently do.