Google AI Series Part 2: Building Your "Second Brain" with NotebookLM
Why the most powerful AI tool isn't about creating more... it's about going deeper with what you already have
This is Part 2 of the “Creator’s Guide to Google AI” series. If you missed Part 1 on Gemini + Workspace, start there.
A few weeks ago, I was trying to refresh my memory on an ads training system I’d used months earlier.
I had already watched the videos & gone through all the docs, so I was really looking for a quickstart. Plus, I’m a visual learner and just needed to remember how all the pieces fit together.
So I uploaded everything to NotebookLM and asked it to create an infographic showing how the entire ecosystem connected.
What came back was a visual map that made everything click. I could finally see how all the pieces fit together—the customer journey, the ad strategy, the content flow. It was like someone had taken all those scattered documents and turned them into a blueprint I could actually follow (I uploaded the infographic below to Gemini & Nano Banana gave it a little ‘Pixar 3D’ touch in my colors - the headline text needs way less shadow, but it works).
That’s the moment NotebookLM clicked for me.
As entrepreneurs, we have a data problem. And it’s not the problem you think.
We have folders full of PDFs we haven’t read. We have 47 open browser tabs of “research.” We have voice notes, webinars, and newsletters piling up in our digital inboxes.
We are drowning in information, but starving for wisdom.
We can create content faster than ever now. But speed isn’t the same as understanding. What if instead of making more, we went deeper with what we already have?
In Part 2 of my Google AI series, we’re stepping away from the “Office” and entering the “Second Brain.” We’re going to look at how Google built a tool specifically to help you learn faster, research deeper, and actually remember what you read.
The star of this show is a tool called NotebookLM, and honestly, it’s one of the most underrated tools in Google’s AI ecosystem.
What Makes NotebookLM Different
NotebookLM is fundamentally different from ChatGPT or Gemini because of one critical feature: Grounding.
When you talk to regular ChatGPT, it pulls answers from the entire internet - which means you get hallucinations, generic fluff, and answers that sound confident but might be completely wrong (Gemini completely missed some of Google’s AI products when I was outlining this series! 🙄).
When you use NotebookLM, you upload your own specific sources—PDFs, Google Docs, website links, even YouTube videos. It then answers questions based only on that material. It can also conduct deep web research for you, looking for new sources (one of NotebookLM's latest features).
No internet, no guessing, no hallucinations, just your sources.
It’s like hiring a genius research assistant, handing them a stack of 10 books, and saying, “Read these and tell me the three most important themes.”
It won’t go rogue and start quoting Wikipedia. It will only work with what you gave it.
How Creators Are Actually Using It
Let me show you three ways people are using NotebookLM that go way beyond “ask it questions about a PDF.”
1. The Visual Mind Map (Infographics)
This is what hooked me.
NotebookLM can take all your sources and create an infographic or visual mind map showing how everything connects.
Upload your research documents, training materials, or strategy guides, and ask it to “create an infographic showing how these concepts connect.”
Why it works: If you’re a visual learner like me, seeing the relationships between ideas is infinitely more helpful than reading linear text. It’s like getting a bird’s-eye view of a complex system.
I’ve used this for:
Understanding multi-part training systems (like that ad ecosystem)
Mapping out this entire Google AI series to see how all the tools connect
Creating visual summaries of research for AI SEO strategies
2. Audio & Video Overviews
If you think AI voices aren’t up to par yet, you haven’t listened to a podcast or video in NotebookLM.
Use any of the sources you’ve uploaded, click one button, and NotebookLM will generate a stunningly realistic discussion where two AI hosts talk about your content.
Not read it to you. Discuss it.
They banter, and they ask each other questions to help you understand the topic. They break down complex concepts into plain language. It sounds like two podcast hosts just finished reading your document and are now having a conversation about it.
AND… ready for this? You can join the hosts in the conversation!
Audio Overview: Perfect for learning on the go. Listen while walking the dog, doing dishes, or commuting. It turns passive reading into active listening (yep, you can download the audio!).
Video Overview: It’s like a great-looking slide deck with a narrator - you select the style, and it does the rest.
Use it as B-roll or background content in your own videos
Share it on social media with visual context
Repurpose it as educational content for your audience
Watch it like you’re sitting in on a conversation (some people learn better this way)
*Here’s an example video 👇🏻. This was from the Google AI Workspace Studio, where I asked what the best Workspace AI Studio use cases were for a solopreneur.
I’ve used the audio version for topics I want to digest in shorter chunks, content I didn’t have time to read, and some YouTube videos.
The video version? I’m experimenting with using clips as social media content and as supplementary material for my courses.
It’s surreal. And incredibly useful for both learning AND content creation.
3. The Content Remix Engine
Try this: Upload your last 5-10 newsletters or blog posts. Then ask NotebookLM:
“Based on these posts, what are 3 gaps in my content that I haven’t covered yet?”
“What topics do I keep circling back to?”
“Create a summary of my core messaging themes.”
It knows your voice because you gave it your writing. It’s not pulling generic advice from the internet; it’s analyzing your patterns, your language, your expertise.
This is how you find the themes and patterns you’re too close to see (and you’ll probably come up with some new ideas, too!).
Why “Source Grounding” Changes Everything
The big LLMs like ChatGPT and Claude are incredibly capable; they know a lot about a lot (how’s that for helpful? lol). But they’re trained on the entire internet, which means their answers are inherently general.
Ask ChatGPT about marketing strategy, and you’ll get solid advice that could apply to anyone running a business.
Ask NotebookLM the same question after uploading your customer interviews, your sales data, and your past campaign results… Now you’re getting insights tailored to your actual business.
Same question, completely different quality of answer.
The difference is context.
Generic AI trained on everything gives you generic answers. NotebookLM trained on YOUR sources gives you personalized insights specific to your situation.
*Sidenote: Obviously, the prompt and input will dictate the output as well, but still, this is based only on your sources.
This is how we move from “Artificial Intelligence” to “Augmented Intelligence.”
It’s not replacing your brain. It’s giving your brain a research assistant who’s already read everything you haven’t had time to read yet...and remembers all of it.
When NOT to Use NotebookLM
Let’s be honest about the limitations.
Don’t use NotebookLM when:
You need it to search the broader internet. NotebookLM only knows what you upload. If you ask “What’s the latest trend in AI?” without uploading recent sources, it can’t answer. Use regular Gemini or ChatGPT for that (although, with the new ‘Deep Research’ option...).
You need image generation like Gemini’s Nano Banana. NotebookLM creates infographics and visual mind maps from your content, but it doesn’t generate creative images like portraits, logos, or marketing visuals. For that, check out Part 1, where I covered Gemini’s Brains & Office, or my post on Google Gems.
You’re working with real-time data. If you need live stock prices, current weather, or breaking news, NotebookLM isn’t the tool for you. It works with static sources you upload.
You want to build automation. NotebookLM is fantastic for research and synthesis, but it doesn’t connect to APIs or other tools (yet). For that, you’d want something like Google Workspace Studio (which I’ll cover in a standalone deep dive soon).
You need quick answers to simple questions. If you just want to know “What’s 15% of $2,000,” don’t upload a source to NotebookLM. Just use a calculator or ask Gemini directly.
NotebookLM is a specialist tool. It does one thing brilliantly: helping you understand large amounts of information you’ve already collected.
Your Action Plan: Build Your First Notebook
We’re not just reading about tools, we’re using them. Here’s your homework for Part 2:
First: What Should Your First Notebook Be About?
Pick something you’re actually working on where NotebookLM’s strengths will shine (though I’ve created a notebook with a single source when I wanted an infographic or mind map of a single topic quickly).
Good first notebook topics:
Learning something new. Create an “AI SEO” notebook and add every article, video transcript, and guide you find. Ask NotebookLM to summarize the fundamentals and give you a starting point. Build your knowledge base as you learn.
Going deeper into a project. Working on a course, book, or big content series? Upload all your research, outlines, and drafts. Ask it to find gaps, identify recurring themes, or create a visual map of how everything connects.
Refining your message. Upload your About page, lead magnets, best blog posts, and client testimonials. Ask NotebookLM what makes your approach different, what themes keep appearing, or what your core message really is.
Understanding a complex system. Learning a new tool, platform, or strategy? Upload all the training materials and documentation. Ask for an infographic showing how the pieces fit together.
My examples:
A notebook for this entire Google AI series (all my drafts, research, and notes)
An “AI SEO Learning” notebook where I add resources as I find them
An ad training notebook with all my worksheets and strategy docs
Pick something that matters to your business right now. That’s where NotebookLM becomes genuinely helpful, rather than just a “fun to use.”
Step 1: Go to NotebookLM.google.com
Sign in with your Google account. It’s free.
Step 2: Create a New Notebook
Click “Create” and name it something specific.
Examples:
“AI SEO Learning Hub”
“Book Project Research”
“Brand Messaging Deep Dive”
“Course Content Planning”
Step 3: Upload Your Sources
Start with 3-5 sources related to your chosen topic. You can add up to 50 sources per notebook, but start small.
Upload:
PDFs, Google Docs, or Word files
Website URLs
YouTube video links
Audio files
Step 4: Ask It Strategic Questions
Try questions that help you understand the big picture:
“What are the main themes across these sources?”
“Create an infographic showing how these concepts connect.”
“What am I missing? What gaps do you see?”
“What should I focus on first?”
Watch how it references specific parts of your sources in its answers.
Step 5: Generate an Audio or Video Overview (Optional)
Click either the “Generate Audio Overview” or “Generate Video Overview” button.
Wait about 2-3 minutes (depending on how much content you uploaded).
Then listen to (or watch) two AI hosts discuss your content like they’re hosting a podcast about it.
Try both formats and see which one resonates with how you learn—or which one you might repurpose as content.
Step 6: Ask for the Infographic
This is my favorite feature.
Prompt: “Create an infographic showing how all these concepts/ideas/strategies connect.”
The visual it creates might completely change how you understand your topic.
That’s it. You just built your first “second brain.”
Next up: Part 3 - The Creative Studio (It’s looking like it’s going to be two parts, so Part 3A & 3B. There’s too much to cover in one post!).
We’re diving into Google’s creative generation tools. The ones that help you make original visual and audio content without hiring designers or spending hours searching stock libraries (I’ve stopped using any stock images since ImageFX & Nano Banana!).
Image generation: Imagen 3, ImageFX for experimentation, and Nano Banana Pro (2) (Gemini’s image generator that’s become my go-to for Pixar-style characters).
Video generation: Veo for creating video from text prompts, Google Vids for turning documents into presentations, and the experimental tools that are pushing what’s possible.
Audio & music: MusicFX for creating royalty-free background music and soundscapes
Specialized creative tools: Whisk (image remixing), GenType (custom typography), and other experimental features in Google Labs
If you’ve ever struggled to find the right image for a blog post, needed custom visuals that don’t look like stock photos, wanted a custom character, or spent 20 minutes searching for royalty-free music that doesn’t sound terrible, Part 3 is going to change your workflow!
Read more about NotebookLM 👇🏻
Daria Cooper - “I Blew My Mind” newsletter. She figured out how to make NotebookLM write in her voice. Worth reading if you want to push the tool further.
This is incredibly helpful, and I agree with others that the “What NOT to use” section makes it even more valuable.
I am probably in the minority, and everyone has different learning styles, but I absolutely hate infographics. I see one it just feels overwhelming unless it’s very simple. I comment on this only to remind people to offer their content in a variety of ways to accommodate those different ways people like to consume information.
Incredible breakdown, Kim! NotebookLM’s approach to turning scattered resources into a “second brain” makes research, learning, and content creation so much more actionable. The visual, audio, and analytic features are game-changing for deep understanding.
I talk about latest AI trends and insights. Do check out my Substack, I am sure you’ll find it very relevant and relatable.