Google NotebookLM Tutorial - Build a Research Notebook That Cites Sources

Key Takeaways

In this module, you will go through a guided exercise where you will build your first notebook, upload multiple types of sources, [music] and organize them effectively. Then, you will create your personalized workflow design that shapes your notebook LM to fit your specific role, whether you're a student, researcher, or professional. Lastly, you will learn how sharing and evaluating can actually help you get constructive [music] feedback and refine your approach. By the end of this session, you will have a complete notebook workflow that you can use right away in your own projects. So, let's start with your guided exercise and build your first notebook. Pick a topic that is relevant to your work or studies. Here, we're going to choose AI in healthcare. So, the next step is to gather your sources. You can collect two or three documents, whether they're PDFs, Google Docs, presentations, or Excel files, and bring them all together in one place on your local system. Then, we create our first notebook. In the left panel, click on new. New notebook and upload your sources. We'll also pull sources using the discover button. We can generate outputs using the studio panel, where we can summarize our sources, chat with them, ask targeted questions, and generate study [music] guides, which notebook LM will organize for easier review. We can create an audio or video overview, where we can choose the tone and length, then preview the result. Pause after each step and take a moment to see if the audio or video overview is correct. To explore the citations that notebook LM provides, clicking [music] through ensures you can see where each insight came from. Now, let's look at your notebook LM interface to see how you can build this AI in healthcare. We'll click on create new notebook. Here, we'll upload our sources. I've clicked on choose [music] files. I've successfully downloaded a few relevant case articles and academic papers concerning the application of artificial intelligence in the field of healthcare. Now, in addition to this, we will proceed to click on the prominent discover option, then carefully select the web category, subsequently type in the specific search query AI in healthcare, and finally, submit our search request. So, while the sources we've uploaded from our local files are being added, we've also gathered sources from discover. Now, based on the local sources you've collected, check to make sure they don't overlap. If something overlaps, deselect it, or if none of them match, don't select any sources. So, let's scroll down and see if these are helpful to us. Now, overall, if you scroll down, none of my collected sources overlap and all of them are relevant. Now, let's import all these 10 sources. By default, the discover option always gives you only 10 sources, >> [music] >> regardless of your plan, whether it's free or pro. Let's import them. Now, the discover sources are added along with your uploaded sources and are organized in the panel on the left side. So, they're being processed now. Once they're processed, [music] you'll see this. When they're processed and added to your source list, [music] you'll see a check mark here. So, >> [music] >> when you scroll down, you'll see this source. It says red, which means it's not being added because it's an invalid URL. You can remove this source to make sure your answer or the study material generated is relevant and doesn't include noise or hallucinations. Now, if you look here, you have a summary based on these sources. So, it says, based on the 17 sources you've uploaded, you have this summary, right? And the name of your notebook is given as regulating AI in healthcare, gaps [music] and recommendation. Now, let's take a look and generate some outputs. For this case study on AI in healthcare, let's start by chatting with our sources. So, if I enter a prompt or a question here, notebook LM will generate an answer based on my input. My input here is, I want to know the main barriers to AI adoption in hospitals. Now, based on my input, it will provide the output. Notebook LM has generated the main barriers, listing several barriers here along with the reasoning. As you look at the reasoning, you'll notice these numbers, which are citations. [music] If you scroll, you see the source. However, [music] if you want, you can click, and here in your left panel, it will show you which source it pulled from and what context it used within that source. It highlights the content from which it took the answer. So, this is how you can chat with your sources in the chat panel. Now, if you want to use the studio panel, for example, if you want to generate a report, as I mentioned in the previous module as well, depending on your use case, >> [music] >> there are suggested formats for reports. In the previous example that was related to market research, here, because our case study is about AI in healthcare, it suggests options like a white paper, a policy briefing, a concept explainer, or an article. Study guide, briefing document, or blog post are common options. So, let's [music] see. We've already looked at the study guide, but let's generate one here as well. And while we're [music] at it, let's see if we can generate a white paper, too. What does it look like? Again, you're free to explore and decide exactly what you want based on your use case. [music] These reports will give you guidance. You can learn about your use case, [music] in this case, specifically AI in healthcare. So, while these reports are being generated, let's also have our audio generated. These generations might take some time, depending on how many sources you've uploaded and how large [music] those files are. That's because notebook LM will go through each of the sources and build these insights for you. Let's take a look at the study guide that's been generated for us. As we saw in a previous example, study guides can include quizzes, short quizzes, long quizzes with answer keys, and essay questions as well. So, as you can see here, everything in your study guide is based on your sources. You can also give your feedback on whether it was a good report or a bad report. For example, [music] if you select bad report, you'll need to explain why. You can provide your input as well. If you say it's a good report, that feedback gets recorded. Now, let's take a look at the report that's been generated for us. The white paper that we clicked for, here you can see it's titled realizing the promise of AI in clinical diagnostics, a strategic framework for trust, integration, [music] and value. All of this was generated based on the 17 sources we uploaded. It has an introduction. It covers transfer, transformative potential. Every point is addressed based on the sources we provided. Again, if it matches your requirements, you can copy, format, and use it. If it doesn't, you can also delete it. So, the next thing is, we generated an audio file. Here, you can play this audio and listen to see if it matches what you want. It's interactive. That means it is. It will ask you questions. >> [music] >> It can also be in a podcast format. So, you review the video, you can download it, change the playback speed if you want, or if you want to share it, you can share it. So, these are the outputs, the intermediate level, advanced level, or basic level outputs that can be generated based on your sources. Here, we used AI in healthcare as a use case and generated these sources. Here, a mind map won't help because this is more of a research topic rather than something to be learned. Mind maps usually work when you want to learn a concept or for educational purposes. Also, if you look here in the prompting area, you'll see that notebook has already suggested some questions you can ask based on your sources. So, if you want, it can explain to you what the Europe AI Act is because we've uploaded a document related to the Europe AI Act, and it gives you the details. Again, if you want to see the citations, >> [music] >> you can always scroll or click on the citations. It will take you to the left panel to the source and show you where it pulled the information from. Now that you've built a sample notebook, let's see how you can design a workflow that's tailored to you. What are the best practices? First, define your persona. [music] Are you approaching this as a student who needs flashcards and a study guide? A researcher comparing multiple studies? Or a professional producing a client briefing? Your persona shapes the outputs you prioritize. [music] Next, map tasks to outputs. For students, [music] this might mean lecture notes, study guides, and quizzes. For researchers, it means papers, contradiction spotting, and synthesized reports. For professionals, it could be slides and URLs, briefing documents, or video overviews. Third is setting up a review schedule. Decide when you will update your notebook. Weekly uploads for ongoing projects or monthly summaries for long-term research. Regular maintenance ensures your notebook stays relevant. Note that improvements help you identify what worked well. Naming conventions, upload order, or discovering additions and places where you could streamline. This step ensures that Notebook LM adapts to your needs, not the other way around. Finally, share your notebook with a peer, teammate, or classmate. Ask them to review the clarity, completeness, and relevance of your materials and outputs. Encourage constructive criticism. Are any important perspectives missing? Are any sections too detailed or unnecessary? Is the study guide or report easy to follow? Use their feedback and your own reflections to refine your notebook. This reflects real-world situations. In research groups, classrooms, or teams, [music] your work often gets better through collaboration and outside perspectives. Don't be afraid to update or reorganize materials based on what you've learned. That's your continuous improvement, and it's part of an effective workflow. To wrap up module four, you built a complete notebook using diverse sources and explored multiple outputs, such as summaries, [music] chat, study guides, and audio or video overviews. Next, you saw how you can design your own personalized workflow to align Notebook LM's features with your own role and responsibilities. We also saw how sharing your notebook and receiving feedback is essential and will refine your approach for real-world use cases. With these skills, you're ready to move on to our final module, which covers best practices and next steps, where we'll focus on accuracy, privacy, and sustainable workflows. If you enjoyed this video, you can enroll in Great Learning Academy for free. Choose from hundreds of courses across multiple domains and earn a certificate of completion along the way. And if you want to take it a step further, you can try Great Learning Academy ProPlus with a free trial. It provides access to additional premium courses with more advanced content from distinguished faculty, along with features like guided projects, a resume builder, and mock interviews to make your learning journey even richer. [music] Link is in the description.