Boosting your tech watch and way of learning by combining Perplexity.ai and Google NotebookLM

Warm up, let's frame it

As 2025 unfolds, it’s already proving to be a hectic year filled with an overwhelming amount of tech news. For tech enthusiasts like me, this often leads to frustration rather than enlightenment.

My goal isn’t just to observe technology trends; I want to understand how they work. However, the time required for this exploration has become increasingly challenging to manage.

In my previous writings, I've discussed how AI and generative AI have streamlined my tech scouting and analysis. Recently, I've discovered a powerful combination: Perplexity (https://www.perplexity.ai) and Google NotebookLM (https://notebooklm.google.com).

Together, they significantly enhance my efficiency in researching and understanding technology.

In fact, I use Perplexity to find the most effective and relevant sources, sources that I use to feed a Google NotebookLM, which is my workshop for dynamically exploring and producing my learning material, and as a bonus I can generate an audio file.

Focus on Google NotebookLM, and hacks

As a reminder, Perplexity is something I could describe as "the search engine I've always dreamed of". Perplexity AI is an advanced conversational search engine that uses large language models to provide accurate and contextually relevant answers to user queries. Instead of presenting a list of links, Perplexity provides synthesised answers from multiple sources, saving users time and effort.

Here, my goal is to set a curated and trusted (as much as possible) sources to feed my exploration and content production.

Each answer includes citations to the original sources, making it easy for users to verify information. In addition, Perplexity supports multi-modal functionality to blend user material and enhance the search experience.

To optimise your searches, and the relevance of sources to feed into the production phase of your intelligence analysis, it's important to understand how some rather structuring operators work, such as the ‘after :’ operator, allowing you to specify a date to retrieve information published after that date, or the ‘filetype :’ operator, allowing you to search for specific file types. Of course, you need to be careful about the effectiveness and real impact of these operators.

As an additional hack, you can use the 'Spaces', which allow you to organise your thematic searches, to specify different types of source, including YouTube videos, which are very useful. To do this, however, you need to activate the 'Pro' option, which you can use up to 5 times every 4 hours with a free account, and with unlimited access if you have a subscription, so it's clearly best to optimise your searches. In the end, all you have to do is copy the search results (the icon is at the end of the answer) and extract the sources to add them to Google NoteBookLM.

Focus on Google NotebookLM, and hacks

Google NotebookLM is a tool to help you understand and explore complex documents, make new connections between information and write your first draft faster. I guess the current version is based on Gemini 1.5 Pro. For me, NotebookLM is a real personal research and writing assistant for exploring and carrying out my regular monitoring and, above all, for delving deeper into subjects by providing relevant sources as input from the stage described above.

The sources supported are quite interesting, with resources coming from the Google drive or be imported from your workstation (PDF, .txt, Markdown, audio like MP3), but also from web links or YouTube videos, or even raw material that you can copy. Be careful, of course, that you remain responsible for the sources and documents you upload, and above all that you don't upload professional documents, which would certainly violate your company's AI usage policy. In my case, it's about watch, so it's essentially public and online information. These sources allow NotebookLM to base its responses on the information we are most interested in.

The "Discussion" section allows you to interact with the sources and drive the exploration of the topic. For example, if you declare video and web sources, the generated content will be annotated with the reference source.

More than just icing on the cake, the "Studio" section automatically generates a summary document that can become a new source, a chronology, a study guide, a FAQ or even an audio summary where you can set the parameters for what the AI hosts should focus on, or target the type of audience, bearing in mind that there is a 'beta' feature for direct voice interaction.

In my case, my experiments lead me to use the discussion as an exploration thread, the study guide and quiz to check that I've understood the subject, and the summary document as an additional resource. I often keep only the FAQ generated, which contains the essentials I need to remember.

Of course, you have to keep a critical eye on the results, and there are two limitations to bear in mind: a notepad can contain a maximum of 50 documents, and each document can contain up to 500,000 words, so 25 million words of source material, plenty of room for exploration.

Outcomes

This duo is transforming my approach to technology monitoring and content production. While the process remains grounded in classic research principles—search, filter, clean, and feed—the results are significantly enhanced by the quality of the sources I utilize.

Moreover, this combination has the potential to revolutionize learning and training. By leveraging the right sources, we can create dynamic learning materials tailored to our exploration needs, paving the way for innovative educational practices.

What tools do you use to streamline your tech watch and training processes?