More often than not, everyone has a collection of 200 PDFs, random Google Docs, and meeting transcripts that never actually get used. It's easy to take notes, but then actually putting these notes into action can be a whole other ball game. Traditional note-taking apps like Evernote and Notion are like digital filing cabinets. They can store all of your information, but they don't actually help you think.

NotebookLM has completely shifted my workflow from "where did I save that?" to "what is the relationship between these three things?" Rather than taking notes and storing them away to never be seen again, I can use NotebookLM to connect dots across my documents that I otherwise never would have noticed.

Why NotebookLM is different

There are tons of AI tools out there, so why this one?

NotebookLM sets itself apart from other LLMs like ChatGPT or Google Gemini thanks to its retrieval-augmented generation (RAG). It anchors the AI specifically to your data, meaning you get no hallucinations, just logic. A downside of this can be that you don't get information from the internet, but in turn, you can be 100% certain that the information it's retrieving is based entirely on your documents.

With a 50-300 source per notebook capacity based on your plan, you can upload entire projects or years' worth of research for NotebookLM to be grounded on. This allows for macro-level synthesis across hundreds of documents that you otherwise may not refer to. Once it has access to all of these documents, you can ask it to retrieve information, analyze your documents or even compare data.

As it feeds you the answers to your questions, it provides you with a citation which links out to the exact sentence within your documents that it pulled the information from. You can provide it with things such as YouTube's transcript, a technical white paper, or even an export of Slack messages. It can then connect the dots between the three and provide you with answers to questions that you otherwise would have to trawl to find yourself. If you find a reliable source online then you can even feed this to NotebookLM and still be sure that it's not engaging with unreliable websites and hallucinating.

I provided NotebookLM with access to all of Shakespeare's complete plays. I was then able to ask it to find common themes. It's worth noting that when providing it with this much information, it does take a while to process, with an approximate 30-second wait time for my queries, which can feel significantly longer than some LLMs' instant response time. When it did finally respond, not only did it highlight common themes like I asked for, but also broke down why this imagery is reoccurring, as well as providing quotes I could use directly from the plays themselves that I could use to potentially back up these points.

I also created a new notebook to analyze some of my content creation scripts which it was able to pull directly from my Google Drive without me having to download and reupload each one. I was able to ensure I was covering all the details I was aiming to within my video scripts by asking NotebookLM one of the predetermined questions it produced. It ensured me that I was showcasing both aesthetic customization alongside the functional performance of the products I was covering in my videos and pulled out direct quotes of where I did this, reassuring me that I was providing a balanced argument.

NotebookLM has been a fantastic tool to analyze my documents without the worry of background noise from the internet, which is often an issue with other LLMs. I don't have to worry about hallucinations and unreliable sources finding their way into my summaries.

Look beyond the chatbot

There are so many hidden features you might miss

There are a range of other great tools on offer beyond just the chatbot itself. Another great feature that NotebookLM provides is the ability to turn your documents into a podcast. NotebookLM provides you with AI hosts that can turn your notes into a debate, which you can listen to even when on the go. It's not just a summary, but goes in depth about the information that you've uploaded, which can often surface contradictions in your own notes. There are a range of other useful tools you can generate in the NotebookLM studio too, including videos, mind maps and flashcards.

Another way that I've been using it is the "everything notebook" strategy. I dump every random thought into a notebook, and it allows me to see what patterns emerge over time. This has allowed me to not just use NotebookLM in a professional manner, but also like a personal diary for myself. While it won't be replacing my therapist anytime soon, it allows me to look at my own thoughts in a more analytical way without having to directly read back over them myself.

Deep Research Mode is a new feature that also allows for an in-depth report and can suggest and provide you with new sources if NotebookLM feels like you're missing something. While it won't automatically add these sources, it can provide them for you to check over, so you can decide if they are reliable enough for you to add to your notebook. This is great because you may not even realize you're missing something exceptionally useful, but NotebookLM will spot this gap for you.

Work or play

I've used NotebookLM for a variety of purposes

NotebookLM has been a truly exceptional tool. It sets itself apart from other LLMs as there is a guarantee that it won't hallucinate. It works off of your sources and your sources alone, so you don't need to worry about random and incorrect information being found on the internet. Not only is it a fantastic research tool, but it can also be used in a variety of other ways, helping you connect dots or identify gaps within your work.