Episode 167 - NotebookLM: An AI Assistant for Shopify Success

Hey, Scott Austin here.

Today I’m diving into a powerful AI tool—Google's NotebookLM—that’s become my go-to for organizing business data, learning from that data, and acting as my digital memory.

AI has come a long way fast. My first episode on AI was back in 2022, and in just a few short years it has moved from an emerging trend to an essential business tool. Now it’s baked into many of the tools you already use—including the Shopify admin itself. These days, I rarely meet a store owner who hasn’t incorporated AI somewhere in their workflow.

With AI evolving so quickly, the real question is: which tools provide genuine value and which are just hype? That’s why in this episode I’m going to focus on NotebookLM, which I find extremely useful and a bit unique in the AI space. With NotebookLM, you’re not just getting another AI tool; you’re getting a dedicated research assistant that can transform your pile of documents, customer feedback, and reports into actionable insights.

What is NotebookLM? An Overview

NotebookLM—short for Notebook Language Model—is like giving yourself a research assistant who never forgets. It’s an AI-powered note-taking and analysis tool from Google that helps you pull meaning out of the piles of data, docs, and transcripts you already have.

How is NotebookLM Different from Traditional LLMs?

Unlike generalist AI chatbots such as ChatGPT, Claude or Google Gemini, NotebookLM operates with a "source-grounded" approach. This is a crucial distinction. What does that mean for you?

Source-Grounded AI
NotebookLM's responses are based exclusively on the specific documents, websites, and multimedia files you provide it. It doesn't browse the broader internet for additional information or just make things up. This significantly reduces the risk of AI "hallucinations" which is that annoying trait where the AI generates factually incorrect content. NotebookLM can achieve up to 94% factual accuracy, a notable improvement over general LLMs. In other words, the quality of insights you receive is directly tied to the quality of the data you provide it.
Privacy and Proprietary Data
Google doesn't use your uploaded data to train its models. This is vital for businesses handling proprietary market research, customer data, personnel files, and internal strategies.
Synthesis, Not Just Generation
While tools like ChatGPT excel at brainstorming and quick copy creation, NotebookLM shines when you need to organize, analyze, and synthesize multiple data files into something useful and usable. It thrives when working with multiple documents, datasets, or detailed reports—helping you digest complex information quickly and uncovering insights that might not be immediately obvious.

So, in essence, NotebookLM offers accuracy, privacy, and deep analytical power that truly sets it apart.

Now that you understand what makes NotebookLM unique, let’s look at how you can begin using it effectively.

For Shopify store owners, NotebookLM can be used to analyze customer reviews, synthesize supplier communications, compare product performance reports, or even summarize internal SOPs. This makes it easier to spot trends, streamline operations, and make data-driven decisions.

Getting Started

Getting started with NotebookLM is straightforward:

Step 1 - Access and Create a Notebook
You can access it at notebooklm.google.com and sign in with a Google account. It has a robust FREE level. You begin by creating a new notebook, which serves as an organized workspace for a topic. And you can have different notebooks for different topics.
Step 2 - Upload Your Sources
NotebookLM can ingest a wide variety of content, allowing you to aggregate content into a notebook. Let's break down some of the most common types you can upload:
1. First, PDFs, Text Files, and Markdown files like research papers, reports, and books.
2. Next, Google Docs and Google Slides (your own or shared documents).
3. Moving on, Web URLs. It analyzes the full content of live webpages.
4. Also, YouTube Video URLs. It uses the video's transcript not the video itself. YouTube makes transcripts for public videos but not for unlisted videos.
5. You can also upload Audio Files. It transcribes and analyzes them.
6. And finally, Copied and Pasted Text.

There's also a "Discover Sources" feature that allows you to search for and add additional relevant sources. This is very useful when you want to incorporate public content into your notebook to give NotebookLM expert content to use in its work.

Using the Data

Once you've created your sources, the fun part begins. And that's getting insights back from the data. There are several ways to do this:

Use 1 - Interact with Your Sources through the Chat Interface

Once your sources are uploaded, the central panel becomes your chat area. You can ask specific questions, request summaries, find connections between documents, and extract quotes. Every response includes quoted citations, letting you trace information back to its origin. You can even select which sources should be used by NotebookLM when responding to you.

Use 2 - Generate Specialized Outputs in the Studio Panel

The right panel offers powerful "Studio" features to process your content. This is the flashy stuff that has gotten a lot of attention. At first, I thought it was more sizzle than steak. But the more I use NotebookLM, the more I use and appreciate these features, including:

Study Guides
Create a complete overview including key concepts, definitions, main themes, essential quotes, and discussion questions. It can even generate quizzes with answers.
Briefing Docs
Provide concise summaries, useful for executive overviews.
FAQs
Generate frequently asked questions based on your documents.
Timelines
Create event timelines from documents containing chronological information.
Audio Overviews
Transform your content into AI-generated, human-sounding discussions that resemble a podcast, which you can listen to to ramp up your knowledge. These can be interactive, allowing you to jump in and ask questions during the "podcast". This may sound like science fiction, and using it feels very futuristic.
Video Overviews
This newer feature creates a video with a slideshow and AI-generated narration, effective for engaging presentations.
Mind Maps
Visualize connections between information from your sources, breaking down concepts and showing relationships.

Use 3 - Note-Taking and Refinement

NotebookLM allows you to save chat responses as notes and even convert these notes into new sources within the notebook, enabling an iterative loop for deeper analysis. You can also add your own manual notes.

Use 4 - Sharing and Collaboration

NotebookLM offers sharing capabilities, allowing you to add other users to your notebook with different access levels, like full notebook or chat only, and include a welcome note. This is particularly useful for team collaboration. More advanced features for sharing and analytics are available in the paid version.

These features, from specialized outputs to collaborative note-taking, truly empower deeper engagement with your data. To make this even more tangible, I want to share some personal examples of how I've been using NotebookLM.

Examples of How I Use NotebookLM

I've been using Notebook for over six months, which isn't a long time. But this is new technology. I want to share three examples of the ways that I've used NotebookLM to help you think of scenarios that may be of value to your business.

1. My Memory Cache

I've got a lot of clients and I have several projects going on for each client at anytime. So my ability to keep all of straight is basically nil. I use Notion to stay on top of the task list. But I find NotebookLM as an incredible tool for actually having a memory I tap into.

I've always used Google Workspace for my agency's email. And that comes with Google Meet which I use for all of my client meeting. I've always recorded my meetings since Google Meet offered that feature. I'd go back to a recording to remember what decisions were made or to create a task list in Notion. Then Meet added text transcription of the meeting conversation. And in early 2025, it added Gemini notes of the meeting.

So, I now create a separate notebook in NotebookLM for each of my clients. In that notebook, I add:

Their website or websites.
Any documentation they have provided me. Especially brand guidelines or strategy docs.
Any relevant YouTube videos they have posted.
The meeting notes or transcript of each of our meetings.
I sometimes add the Notion tasks for the client. But that's a bit of manual work as I need to export the tasks from Notion and then add them as a source in NotebookLM.

Once, the sources are added, I add all team members from JadePuma. We can then leverage NotebookLM to analyze the data and act as my virtual memory.

Here's one recent small but common example. I was working on a campaign for a client where we were offering a 10 percent off campaign. I couldn't remember what past campaign we did that had a similar discount. So I got the answer for NotebookLM and quickly found the old campaign and copied it as the starting point for the new campaign.

I also use the Audio Overviews or Podcast feature a lot more than I thought I would. I find them as a way to reset my view of the brand by seeing the big picture that NotebookLM comes up with. And they are a great way for new team members to get some background on the brand and understand the learnings we've gained through the experiments we've done.

I'm finding this virtual memory super helpful on an almost daily business. So I've ensured that I have processes in place to keep adding sources to a client's notebook as we create new content.

2. Podcast Gap Analysis

I've now done 167 episodes of this podcast. More topics than I can remember. So I don't remember what I've already done and what topics I want to cover in the future.

So, I've created a notebook for the podcast. I've added the YouTube video for every episode as sources. For this, I found a Chrome plug-in that automatically added all the videos in a playlist to NotebookLM. This saved hours of manual entry time. Then I added the episodes from the other top e-commerce podcasts. For this, I found another Chrome plug-in that worked for podcasts.

Next, I had NotebookLM create a list of desired e-commerce topics by looking at the topics the other podcasts have covered. This, of course, is based on the assumption that the other podcasts are delivering the content that e-commerce owners want.

Then I had NotebookLM create a gap list between the content that e-commerce owners wanted and the content this podcast has delivered.

I then pruned the gap list of topics that I don't think belong in the Shopify Solutions Podcast.

Now I have a list of topics to cover in future episodes of the podcast.

3. Product Manuals

Another practical application where NotebookLM has proven invaluable is in managing my short-term rental property.

I have a home that I use as a short-term rental. It rents well and gets all sorts of people coming through it, especially many foreigners who don't know English. And AirBnB translates my text messages with people who speak other languages.

I've created a notebook for the property. For sources I've added:

The electronic version of the manuals for appliances, furniture and more.
YouTube videos I've found on how to use the products in the house. For example, I found a video that shows how to set up the pack-n-play in the house.
YouTube videos that I've created as part of the house manual for guests.

So when I get a question from a guest, I ask that question of NotebookLM. It will give me the best response from sources and cite which one it uses. I then send that response to the guest. I may edit it beforehand. And if they don't speak English, it will get translated by AirBnB.

These are just a few ways I've integrated NotebookLM into my daily work.

Examples: How Shopify Store Owners Can Leverage NotebookLM

Now, let's explore how a Shopify store owner or e-commerce team can practically apply these functionalities to build and grow their business. I won't go into these in detail because the process will be different for each brand. I just want to get you thinking about scenarios that you want to explore. Here are some key examples:

1. Competitive Intelligence

Gather all information about your competitors—their websites, product pages, social media, pricing sheets, news articles, and product demo videos. Upload these to a "Competitive Intelligence" notebook. Ask NotebookLM to compare competitive positioning, product features, and marketing messages, or to identify gaps in their offerings that you could exploit. This instant synthesis helps you spot strategic shifts in the market.

2. Customer Service Knowledge Base & Training

Establish a "Support Team" notebook as a shared resource for your customer service team. Upload company policies, the public-facing FAQ, and detailed product manuals. NotebookLM can generate standardized response templates for common inquiries, and new hires can query the notebook for quick answers, significantly reducing training time.

3. Content Strategy & SEO

For organic growth, create a "Content Engine" notebook. Upload top-ranking articles for your target keywords (as PDFs or URLs) and SEO research reports. Prompt NotebookLM to analyze competitor content for themes, keywords, and structure, then generate a comprehensive, SEO-optimized blog post outline that fills identified content gaps.

4. Market Research & Niche Validation

Use NotebookLM as a "Market Intelligence" hub. Upload industry reports, competitor website URLs, YouTube product reviews, and expert interview transcripts. Then, query it for emerging trends, and market gaps specific to your niche. For example, a query could be: "Based on all sources, summarize the top 5 emerging trends in the sustainable pet toy market. For each trend, cite the source and explain its potential impact on a new e-commerce store".

5. Listen to Your Customers

Create a "Customer Voice" notebook by uploading customer reviews from competitor websites, Amazon or your store. Also add long-form reviews from YouTube.

Ask NotebookLM to identify what features and value props resonate with customers and to extract the language they use. Use this to formulate your messaging to customers. This way, you'll be talking about the things they care about using the words they use.

6. Product Description Generation

Build a "Product Bible" notebook for each product. Include technical specification sheets, high-resolution product images, competitor product page URLs, and relevant customer reviews. Also, upload a document with your e-commerce copywriting best practices. Then, use persona-driven prompts to generate unique, benefit-focused product descriptions that convert visitors into customers.

You can use this to create different landing pages for different personas or ad targets.

7. Proactive Customer Support & FAQ Generation

To reduce support tickets, create an "FAQ Generation" notebook. Populate it with customer support tickets, product reviews (especially those with questions), social media comments, and product manuals. Ask NotebookLM to identify the most frequently asked questions and then generate clear, concise answers, creating a data-driven FAQ page that addresses real-world concerns.

8. Repurposing Content for Marketing Channels

After creating a well-researched blog post or recording a webinar, upload it to a "Content Repurposing" notebook. Add documents with channel-specific guidelines (e.g., LinkedIn, Instagram, email). Then, prompt NotebookLM to extract key takeaways and generate platform-native content like tweets, LinkedIn posts, Instagram captions, or email newsletters, maximizing your content's reach and impact.

Accelerate the Learning Curve

These eight ideas are just the tip of the iceberg, and you'll undoubtedly discover more as you use it. So that brings us to our next area of focus: mastering the full potential of NotebookLM and integrating it into complex workflows. While it's excellent for organizing, analyzing, and synthesizing multiple files, its purpose differs from quick brainstorming tools like ChatGPT. Advanced scenarios can be time-consuming.

Tips to Accelerate the Learning Curve

To make the most of NotebookLM and quickly become proficient, consider the following key tips:

Start Simple and Expand Gradually
- Begin with a single, clear use case and a small set of data or documents. For your first attempt, upload a PDF you've been meaning to read or an article relevant to your work.
- Test its capabilities, such as asking for product recommendations, and then gradually expand your use of its features.
Understand its Core Advantage: Source-Grounded AI
- NotebookLM's primary strength is its Retrieval-Augmented Generation (RAG) approach, meaning its responses are exclusively based on the sources you provide, truly embracing that 'source-grounded' nature we discussed earlier. This significantly reduces the risk of AI "hallucinations" and ensures factual accuracy.
Diversify Your Sources
- You should give your notebooks multiple complete, expert data sources to use. NotebookLM supports a wide range of input formats. Bring in different sources in different formats
- Utilize the "Discover Sources" feature to find and add additional relevant web articles or PDFs related to your topic.
Master Prompting and Asking Targeted Questions
- Engage in conversations with your content by asking specific questions, requesting summaries, identifying connections, and extracting quotes.
- If responses are too generic, refine your questions, refer to specific sections or pages, and provide clear context for the desired output.
- NotebookLM can also suggest prompts if you're unsure what to ask.
Save and Organize Your Work
- Always save any valuable answers or insights as notes, as NotebookLM does not automatically retain chat history across sessions.
- You can convert these notes into new sources, allowing for layered analysis and refining your content.
- Rename your notebooks with descriptive titles to keep your work organized and easily searchable.
- Labeling sources with a topic like Review, Research, Scorecard, etc. will help you more quickly turn sources on and off as needed.
Utilize the Chrome Extension
- Install the NotebookLM web importer Chrome extension to quickly add web pages or YouTube videos as sources to your notebooks while browsing.

By implementing these strategies, you can significantly accelerate your learning curve and leverage NotebookLM's capabilities to streamline your research, content creation, and strategic planning.

One Final Consideration

OK, I just spent a whole bunch of time explaining all of the wonderful things that NotebookLM can do for you. But here's the thing that I want you to understand, even though I called it your research assistant, I recommend that you don't think of it as an employee or a person. Instead, think of it as a tool. Many store owners that I talk to expect that they'll click a few buttons in AI and maybe give it a couple of pieces of information and then the awesome power of AI will deliver them the final product that they are looking for. But it doesn't work that way. NotebookLM, and AI more generally, is just a tool in your workflow. A super-powerful tool, but still just a tool. You are going to need to provide it the information and inputs. You are going to need to map out the workflow to follow. Some of the steps in the workflow will probably be done by you. And, most importantly, you'll need to judge if the output is accurate and meets your needs.

Let me illustrate this point with an example. I used NotebookLM as a tool to create this podcast. Here's the process I went through:

I used NotebookLM to discover sources that talked about NotebookLM. And I also used it to find sources that stressed using it it e-commerce. I added the sources to my notebook. But I went through all 40 or so of them to check if they were appropriate. My personal review removed about 10 sources.
Then I had NotebookLM create a podcast episode that I listened to. It was a useful summary for me of the data I had collected. But it didn't make a good podcast episode for you. So I then gave NotebookLM some detailed guidance on how I wanted the podcast to be structured. That was much more useful. But I knew it wasn't appropriate for the final transcript. It wasn't in my voice and the example were a little too pie in the sky.
So I took that guidance that I gave the audio podcast and instead asked NotebookLM to create text transcript.
I took the text transcript and formatted it so that it was more readable for me.
Then I rewrote the entire transcript to be what I wanted this episode to be based on all that I learned during this process.
Then I had ChatGPT evaluate the transcript and provide feedback. I evaluated all the feedback and incorporated some of them into the final transcript that is this episode.

Hopefully, this example illustrates what I mean when I say that NotebookLM is a tool and shouldn't be treated like an employee. Don't get me wrong, the more you can automate with AI, the better. I've just found that it's good to keep a tight control on the process and frequently use your judgement to check the work done by AI.

Conclusion

So, with that caution in place, I still believe that NotebookLM can provide you with meaningful help. In essence, NotebookLM is not just about automating tasks; it’s about making smarter, more informed decisions by turning your scattered data into a cohesive, queryable intelligence asset. It acts as a dedicated "second brain" for your business, allowing you to operate with better strategic depth and analytical rigor.

And NotebookLM is one of those tools that has many possible uses. So as you are using it, you'll probably come up with new ideas you'll want to test it on.

If you haven't tried it yet, you should. Try uploading your recent customer reviews and ask it what your shoppers love most—it's a great starting point.

Thanks for listening.