This is the year Google is going all in on AI, both Gemini 2.5 and NotebookLM just got some huge upgrades, especially Gemini 2.5 Pro, which is definitely an underrated AI models. It's fast, super smart, and so powerful. So in this video, I'll share how you can combine the two most powerful AI from Google right now to do your work much faster. One of the most underrated capabilities of Gemini 2.5 is its superior coding capabilities using its canvas mode with its enhanced reasoning and thinking. And to my surprise, it is really fast. It can literally build anything which is ready to ship and is fully functional. Unlike on Claude 3.7, you might still need to set up the environment to make the codes really working, so it will be a good choice for non-technicals. And also it has strong multi-model ability, has 1 million token context window and Google is planning to expand to two millions tokens. Just make it a perfect tool for doing different types of complex tasks. It can now assess YouTube videos both on Gemini and AI Studio, and you can ask any questions. Gemini 2.5 Pro is now free for all users without any paid subscription. You can just access it on both Gemini app or through the AI studio and not just Gemini 2.5, NotebookLM's discover sources is incredibly useful. You can just type anything and it will evaluate hundreds of potential web sources and select the most relevant and credible ones for you all powered by Gemini So we can use NotebookLM to curate high quality and verified sources with the least hallucination, extract quality insights, and use Gemini 2.5 as the powerful execution engine for building anything without coding skills. The first one is interactive visualization. Given the 1 million tokens and native abilities to process different content formats, Gemini 2.5 is absolutely incredible at handling super long documents with images and generate interactive visualization. So combining with NotebookLM's insight from verified sources, you can transform any insights into beautiful visualization or tools without any coding skills. First, let me show you how powerful Gemini 2.5 in analyzing complex documents. I have this long patent documents, everything is in images, and it is full of complex diagrams. Now go to Gemini, select 2.5 Pro. Make sure you turn on canvas as we are going to build a visualization. Then upload the patent document and use this prompt to ask it to summarize this document and visualize the key process . Very quickly, It will summarize the whole document with the summary and visualize the key process. I have tried the same prompt on Claude or ChatGPT and it's not working or with enough details given the complexity. And you can always share the canvas and you can see it is mobile responsive, which is very thoughtful from Gemini. Now we can make this workflow even more useful to add export functionality. This time, let's start with Gemini Deep Research with the Gemini 2.0, and you can also generate audio overviews within Gemini. Now export this deep research into word document and download it. If you want, you can always import it to NotebookLM as a source. Again, upload this to Gemini 2.5 and ask it to turn it into interactive infographics. Sometimes there may be missing icons and you can always ask it to fix it. Immediately, we have a really decent infographics about the report and you can hover to show the details and you can also build all the visualizations like micro-sites or web presentations. Now we can ask it to add a download as PDF button that is fully functional and quickly you can see a download as PDF button is added. But note, you need to copy the code and run locally or your hosted environment as Google will block you from running any code using the Canvas link. What amazing is that this is fully functional without setting up anything even for Claude, it will require you to do some setup and not just for PDF. You can try export as image, SVG file, depending your needs and share with your teammates or clients. Now we can combine Notebook LM to find verified sources using the Discover feature. Create a New Notebook. You can directly click the Discover, or you can click Add and the Discover sources. If you have no idea, you can always click "I am feeling curious" to get some random topics. Let's say we are doing research about EdTech industry. We can specify like research reports about core industry trend and NotebookLM will suggest at most 10 sources at a time with a blurb and will always prioritize authority sources from reputable brands, and you can always ask it to focus specific sources like YouTube video sources, big consulting firms or specific brands like Google, Microsoft, or even file formats like PDF, date recency, so definitely play around with this as I found this is so useful to find some quick and reliable sources and note that even it can scan the source doesn't mean you can input all of them successfully. Make sure all sources can be imported and now use this prompt to ask it to identify any key trends and consumer behavior shift. Note. I have asked it not to include any citation, so we can make sure the output is clean for Gemini as input. Then copy the response. On Gemini 2.5, ask it to build an interactive opportunity map with visualization, using the NotebookLM Insights. Within a minute, we have the opportunity map ready. With Gemini labeling each trend, like growing, emerging, established. What I really like is the strategy decision helper. You can just pick the target audience, technology focus, and it will highlight the relevant trend for this target segment with its potential to inspire your thinking. So think about what matters for your business, and you can turn any NotebookLM insights to useful build using Gemini. Before I move on, if you make this far into the video, you're likely excited about the powerful combination of Gemini and NotebookLM. That's why I recommend you also check out this free ebook from HubSpot, "The Marketer’s Guide to Google Gemini & NotebookLM". You can find the download link in the description. It's completely free. This guide shows you how to cut your research time from hours to minutes while uncovering deeper insights than traditional methods allows. It walkthroughs everything from automating research using Gemini deep research, building personalized knowledge hubs with NotebookLM, to real world use cases for marketers. This is the great starting point if you want to incorporate AI into your research workflows. What I like is the 11 step-by-step practical application workflows that you can immediately apply, like content creation, marketing strategy, customer research, industry analysis. These show you how to combine AI tools to maximize productivity and insights. I recommend you download this in the description below for free, and thank you HubSpot for sponsoring this video. Next, research to landing page. Instead of just asking Gemini to build the page or website, we can combine the knowledge and insights from NotebookLM about your target audience and competitors. The best thing is that you can always curate your own sources on NotebookLM, making sure they're verified and high quality. Let's say I'm building a landing page for a coworking space business. First on NotebookLM I can use the discover source to identify relevant research study and surveys about coworking space and prioritize PDF reports So it has found 8 sources for me and we are not done yet. Also, import some top competitors in your niche to analyze their messaging and the language that they use. Try to find at least 4-5 competitors. Now ask it to identify the language patterns, voice and tone about the target segments, who are freelancer, remote workers, and small business owners. So now it will based on the sources and give you the language patterns, specific considerations for these target segments as well as the tone. And if you have other own sources, like call transcript, surveys, also input to NotebookLM. And this will be very useful as we build the page on Gemini. Now copy the response and paste to Gemini. Let's say this is a website design that inspires me and I want to adapt it for my business. Then we can take a screenshot and use this prompt to generate the landing page mockup and use the NotebookLM Insights as guidance. So you can see it will not only generate the landing page using the layout I specify, but also the word choices and messaging. You can definitely tell it is using the language and addressing the concern in the membership benefits. And it is beautifully designed and ready to deploy. So make sure you not just ask Gemini to build, but always give it meaningful insight so the output will be much better. Next is building learning materials. Everyone knows NotebookLM is an incredible tool for learning. Now with Gemini 2.5 strong coding ability and NotebookLM's mind map features, we can transform any learning process with powerful visualization and interactive learning materials in just minutes. This is a powerful workflow if you want to learn any new subjects or you need to create educational content. For example, I'm learning or teaching Data Science. So on NotebookLM, we can first find some YouTube tutorials about data science and this is why I like the Discover Source features. You can just specify the format and it will find all the credible sources for you. And then also find some PDF ebook about data science. And I'm not kidding, NotebookLM even helped me to find a whole book also import them as sources. And remove those that can't be imported. First, we can click on the mind map features and it would immediately generate the visual mind map that gives you the holistic view on how different concepts are related in data science. And you can expand each node to expand underneath concept further. And you can always click on each node, and it will trigger a prompt for details. So this will give you a very structured way for learning any new subjects. Now we are going to build a learning materials using this prompt to ask it to design sets of learning modules about data science. Immediately it will break down into different learning modules with details. So this will be helpful whether you are just doing self-learning or teaching, and this is the power of the 2 million token context window of NotebookLM. To make it even more useful, we can now copy the response, paste on Gemini 2.5 and ask it to transform these modules into sets of interactive learning quizzes. Within just a minute, the learning quizzes are ready and it looks amazing. What I really like is how it breaks down the quizzes by modules as specified from the NotebookLM's insights. So every modules we have different self-assessment and it's fully functional. You can just upload to your server, or if you want, you can just share this canvas link and it is totally working fine. Another bonus is that we can actually turn any NotebookLM mind map into an interactive cluster. So let's say we have this learning project about UX design. Again, as usual, we generate the mind map visual and expand all the nodes, and then export it as an image. Now on Gemini 2.5, upload this mind map image and use this prompt to build a topic cluster visualization, showing how different concepts relate within this broad topic and then immediately it will turn the NotebookLM mind map into an interactive cluster. And not just for learning, you can use it to visualize your content strategy pillars, strategic planning priorities, so the use case are endless. The next one is research your product prototype. For most builders, startup founders, product managers, one of the challenges is how to validate ideas quickly based on market needs. Now with NotebookLM and Gemini 2.5, this process will become much easier and you can iterate faster and reduce the risk of building something nobody wants. For example, I'm building a workplace productivity tool. So first on NotebookLM, let's discover some sources around the recent workplace productivity trend. And I like how NotebookLM will always scan sources from reputable and authority brands and government. Let's just import all of them. Now also discover sources about audience pain points with workplace productivity software and focus on real user feedback from forums. You will see it will try to prioritize sources from Reddit and also, import all of them and remove those that cannot be imported. Of course, if you have other relevant sources, like your own research data, also upload here. Now use this prompt to ask it to identify the five unmet needs in workplace productivity tools and the most requested features from users and the recent trends that might impact how we build this product and organize them by priority. So this is what I really like about NotebookLM is because it can always find linkage between all the sources and provide the insights about your market, customers and products. Then copy the response and paste back on Gemini 2.5 as we are gonna build the prototype. Use this prompt to ask it to create a beautifully designed and fully functional product prototype for a smart workplace product that address the insights we gather from NotebookLM. Immediately, you will see the prototype is ready. It's fully functional, and I can see it's also addressing the top requested user features like Ai-powered Assistant, timely tasks reminders, integration with other existing tools, and even a focused time zone that address the target customer needs. So this will be very useful to immediately transform any of your quality research insights into a product mockup that is ready to build. The next one, create podcast script and audio. NotebookLM's audio overview generations is probably one of its signature features and everyone loves its podcast style conversation. And the core technology is powered by Gemini. And using Gemini 2.5 multimodal capability, we can easily transcribe any audio, re purpose it to another podcast script and even generate a ready to use podcast audio track. Let's say I'm building a podcast episode about finding a dream job. As usual, we import all the sources and then we can easily generate the audio overview by clicking the generate button and the podcast style conversation is ready. So download the file. Now go to Google AI Studio. The reason for this is because this is the developer interface and support more file format uploads. So make sure the Gemini 2.5 Pro model is selected and then upload the podcast audio file and use this prompt to extract the full transcript. And you can see this is super amazing and Gemini multimodal capability is really powerful. It extract the full transcript with timestamp. And if you have YouTube videos, you can also use the same approach to get the transcript easily. Then use this prompt to convert this two person podcast script into a first person single host audio script as if I am speaking directly to my audience. So you can see immediately this script is repurposed using first person view. And if you want, you can always make adjustment and then record using your own voice, which is a preferred approach. But if you don't want, you can actually copy the script and go over to the Screen Realtime, a feature on AI Studio that allows multimodal interaction with AI Now pick the output format and desired audio voice. Copy the script, and use this prompt and ask it to read the podcast script word by word and skip any transition effect. Okay, let's dive into this. You're looking to find your dream job , not just any job, right? You've maybe even looked at a career research guide, which tells me you're serious about this. You're not just messing around on job boards anymore. You're ready to really figure out how to land a role.. You can hear that. This is absolutely incredible. And what I noticed is Gemini's audio has improved compared to months ago. You can definitely hear the intonation, pacing is very natural, and if you are satisfied, you can pause, and download it as a high quality audio file. And one tip is try to break it down by different smaller sections as sometimes the connections of Stream Realtime may suddenly be lost and you might need to start over again. And of course, you can also use other text to speech AI tools or have your cloned to AI voice. But using this method, you can turn any script into ready to use audio, whether it is a podcast episode, training materials, professional narration to reports. So there are so many applications you can use this. So you can see from these use cases what really matters is about your source quality because Gemini 2.5 can literally build anything, but only when you give it quality sources, the output will become meaningful. So that's why NotebookLM is so useful in this workflow because you can always curate your own sources and have the biggest control on the source quality. And I really hope Google can seriously consider integrating NotebookLM and Gemini 2.5 Canvas. This will be the real game changer. And if you want more inspiration, join my community. You can find all the tested prompt I share on my channel and other featured content. You can find the link in the description and before you go, also check out this whole playlist of videos about AI to get more use case inspiration. I will see you next time.