I almost gave up on Gemini Notebooks, but then I found this feature

While I’ve long wanted to organize my Gemini chats and files into projects, when the feature arrived as Gemini Notebooks, I initially felt disappointed.

However, after playing around with the feature using both Gemini and NotebookLM, it’s more useful than I expected.

When I envisioned a projects feature in Gemini, I wanted it to be a useful place to organize files and chats into folders. This was based on my experience with Claude, especially its own projects feature, which makes it easy to organize files and chats under a specific folder.

But when Notebooks arrived on Gemini, I didn’t enjoy its NotebookLM integration. It felt like it was just a way for Google to push users to use both products rather than refining each separately.

The fact that your Gemini Notebooks also automatically sync to NotebookLM can be a drawback, since you don’t have unlimited notebooks available for your account.

After using the feature a bit more over a few days, I did come to appreciate its approach to organization — even if it wasn’t quite what I had hoped for. However, it’s when I realized what these notebooks could do for my NotebookLM research that I really came to value the integration of the two tools.

When using Gemini Notebooks, I realized that my chats with the AI were synced to NotebookLM as a source. This may seem redundant at first since NotebookLM also has chatbot features, but it unlocks one powerful tool that NotebookLM doesn’t really have — access to the web.

NotebookLM grounds itself in your sources, meaning that it cannot pull information from the wider web to add context to your responses. This is useful when it comes to preventing hallucinations, but can limit the AI tool’s insights.

Once I realized that my chats with Gemini were being synced as NotebookLM sources, I figured out how to use these chats to add context. NotebookLM can pull links from the web that you can use as sources, but this is a somewhat clunky process that doesn’t always surface the best information.

It also requires there to already be a source on the topic available — which is more difficult to find when it comes to niche context. I’ve also found that it’s not always easy to find sources relevant to your local context if you don’t stay in a Western market.

But with my Gemini chats, I can now add context without needing to add a separate note or source to my notebook. I can also troubleshoot problems with Gemini and have the solution applied to my NotebookLM projects.

Importantly, only chats within Gemini Notebooks are synced to NotebookLM. You can also select which chats you want NotebookLM to use for a response, or delete a specific chat entirely.

The first NotebookLM project where this feature was immensely helpful was my garden care notebook. In it, I store information related to caring for the plants in my garden.

With Gemini, I could add the context that I live in Cape Town, South Africa. I then asked the AI about how often I should water a specific plant in my garden during the summer. It could use this context, including my region’s lack of rainfall during summer, to plan out a watering schedule for the plant.

I also used it to figure out why another plant was experiencing yellow leaves. While most sources mentioned potential root rot, I could add the context about a recent heat wave. The AI concluded that heat stress was the more likely culprit based on my watering habits.

Back in NotebookLM, I could generate a more accurate watering schedule based on my sources and my Gemini chats. This proved to be more in line with my plants’ real-world requirements than the context-free suggestions from the sources. I also finally got a watering schedule for my grass, which no source was able to determine before.

My next project is to create a NotebookLM hub to explore chronic kidney disease in cats, since my cat has recently moved from Stage 1 to Stage 2. Gemini’s ability to easily add context will be helpful, since I will be able to use information about my cat without needing to upload it separately.

While Gemini’s propensity to ask follow-up questions does annoy me at times, it’s also useful for seeing where I might have gaps in context.

Gemini Notebooks aren’t the projects feature I envisioned, but maybe that’s a good thing. After all, they offer something a little different.

The feature isn’t without flaws, as Gemini still has a habit of making mistakes. For example, I spent around 10 minutes arguing with it over its identification of a plant based on several pictures.

Overall, though, it provides organizational features while adding helpful context to NotebookLM with little effort. Though, as with any AI tool, it remains important to check the accuracy of its results.