I'm currently smack in the middle of finals season, and NotebookLM is one tool I can't believe I once survived academics without. It's the one tool where every feature feels like it's earned its place, and each launch feels like it's been shipped by someone who's actually had to cram for an exam at 3 AM. The tool has come a very long way from its initial days though. It's now filled with features and keeps getting better daily.

Interestingly, a good chunk of people still think all the tool does well is turn your PDFs into podcasts. Now, Audio Overviews are great, and I'd be lying if I said it isn't the feature that made me think, "hey, this tool is going to be big." However, if that's the only feature you associate with the tool, you're barely scratching the surface. NotebookLM's Studio panel is packed with features that most users scroll right past, and the one I think deserves way more attention than it gets is Infographics.

Infographics launched alongside Slide Decks

And most people missed it

If you've explored NotebookLM past Audio Overviews in the last few months, I'd be willing to bet you've used the Slide Decks feature at least once. I'd also say that it's talked about and hyped up nearly as much as Audio Overviews, and given the fact that it does a great job at delivering what it's meant to do, that hype is very well deserved.

However, what most people don't realize is that a feature called Infographics launched at the exact same time as Slide Decks. Both of the features dropped in November 2025, and both use the underlying Nano Banana Pro model to generate visuals. Slide Decks got all the buzz, and Infographics quietly slipped into the Studio panel without nearly as much fanfare. Based on Google's own wording, Infographics lets you create high-quality visual summaries of your sources. In their support documentation, Google explains that Infographics are designed to help you understand the main points and visualize data in a fun and engaging way.

As mentioned above, Infographics are generated using Google's Nano Banana Pro model. While this image model did disappoint me a bit when I pitted it against ChatGPT Images 2.0 in certain scenarios, it really shines in Infographics and Slide Decks both.

Now, I think the reason why a lot of people use Slide Decks and not Infographics as much is because the former has a clear-cut use case. You need a presentation, you click Slide Decks, done. Infographics, on the other hand, feel a bit more abstract. Most people see the option and think, "okay, but when would I actually need this?" And that's exactly where I think they're wrong.

When you ask NotebookLM a question via a chat panel, it parses your sources as relevant, and then gives you a text-based answer. When you generate an Infographic, NotebookLM does the same thing, but arranges its findings into a visual hierarchy that makes patterns and connections click in a way a wall of text simply doesn't. For instance, if you want to see how all your sources connect with each other, NotebookLM's Mind Maps feature is clearly the better option. But if you want to distill your sources into something that highlights the most important takeaways at a glance without someone needing to read huge chunks of text, Infographics are the way to go. You could send one to a colleague who doesn't have time to read your full research, drop it into a report for quick context, or use it as a last-minute study aid the night before an exam.

Infographics that actually know what they're talking about

They're actually grounded

Whenever I talk about Infographics, whether it's in an article or a conversation I'm having, one question that's bound to come up is "why not just use Gemini, ChatGPT, or even Meta AI to generate an infographic instead?" Given that all of these tools can generate visuals and don't require you to create a notebook and upload your sources, that's a fair question. But there's one fundamental difference that makes NotebookLM's version far more useful: everything it generates is grounded in your sources. When you ask ChatGPT to make you an infographic about, say, a technical architecture, it's pulling from its general training data. It'll certainly give you something that looks far better than decent, but it is essentially a summary of the internet's general understanding.

With NotebookLM, you won't need to worry about this. It only works with the material you've uploaded, which means the infographic you get back is tied directly to your research, your notes, your documents. Nothing is pulled from the open web, nothing is hallucinated from some generic dataset. That's a massive difference when accuracy actually matters. So, for instance, I like to use Claude Code and NotebookLM in parallel a lot. The former does the coding and the actual heavy work, and NotebookLM helps me understand and visualize what Claude Code actually built. Now, listening to two hosts talk about code wouldn't really do all that much for me. It wouldn't help me make sense of a codebase, and it certainly wouldn't help me understand the architecture of a project. But a visual summary is something that would help.

What Infographics look like with a real project

NotebookLM infographics with real sources hit different

A few weeks ago, I vibe-coded a replacement for my entire productivity stack, and I imported the entire codebase into a NotebookLM notebook. I then headed over to the Studio panel, clicked Infographics, and prompted it with: "A system architecture diagram for 'ProductivityStack,' a local-first productivity web app. Organize it into clear horizontal layers showing how the parts connect." NotebookLM broke the entire project down into five clearly labeled layers: the client-side frontend, the state and communication bridge, the backend server, the local SQLite persistence layer, and the external integrations. It correctly identified specific libraries like TanStack Query and better-sqlite3, mapped out the Vite proxy on port 3001, and even noted that OAuth tokens are stored locally so they survive server restarts.

All of this information was buried in the code, and NotebookLM parsed through all 27 files to surface it. While it isn't perfect (you'll see a typo in the backend layer where automatically is spelled as automaticallonically), the actual architecture is accurate. How the layers connect, what talks to what, which integrations are read-only versus full CRUD — it got all of that right. Making something like this manually wouldn't only have taken effort to design and put together, it would have also required me to go through every single file myself to figure out what connects to what. NotebookLM did both within minutes.

Infographics are NotebookLM's most underrated feature

If you've been using NotebookLM and have never clicked on Infographics, you're leaving one of its best features untouched. It's not flashy in the way Audio Overviews are, and it doesn't have the obvious use case that Slide Decks do. But once you try it with sources you actually care about, you'll wonder why you ever scrolled past it.