[Organizational AI Utilization #062] How to divide NotebookLM units
Hello! This is Terada.
Currently, I serve as the Representative Director of the AI Digital Community (ADC), a community for AI practitioners such as AI promoters at digital-related companies, as well as the Representative Director of FURIKAKE Partners Inc., which supports the 'AI-ification' of client businesses, and AI Portalize Inc., which provides products that support organizational AI utilization. I support organizational AI utilization from various perspectives!
Recently, while providing AI support both inside and outside of companies, I often hear about attempts to use NotebookLM to create a so-called 'internal Wikipedia.' The idea is to compile internal documents, meeting minutes, sales meeting notes, technical specifications, etc., so that AI can answer questions when asked.
In this context, the first thing people struggle with when they actually try to do this is 'how to divide the Notebooks,' and I get asked about this quite often.
Why it is better to divide them into smaller units to some extent
As a result of various trial and error, I believe that it is definitely easier to use if you divide the Notebooks into smaller units to some extent.
For example,
Development Notebook
: Specialized in development-related information such as GitHub documentation, technical specifications, design documents, and development-related Slack logsSales Notebook
: Specialized in information necessary for sales activities such as proposal materials, sales meeting transcripts, sales manuals, and success storiesCommon Notebook
: Information that everyone refers to, such as product overviews, company introduction materials, and company-wide rules
By doing this, if you ask a question in the Development Notebook, unnecessary sales materials will not appear, and in the Sales Notebook, detailed technical specifications will not be mixed in. The AI's answers will become much clearer.
Not just by department, but also by theme
Notebooks for each department are easy to use, but by themselves, they are inconvenient when you want to 'look at things across the board.'
Therefore, it is also useful to prepare Notebooks by theme.
For example...
'Project A Notebook' -> Register development design documents and sales proposal materials together
'Customer X Notebook' -> Register sales meeting records and development support history together
By doing this, when you want to 'see the entire history of this customer,' you can pull it all out at once from a single Notebook.
Align the nature of information for each Notebook
Since NotebookLM is a mechanism that searches for relevant information from uploaded documents, if you put information with different natures into the same Notebook, the accuracy will drop.
For example, it is better to keep sales meeting minutes and development API specifications in separate Notebooks.
Conversely, it is fine to group items that share the same usage scenario, such as sales records and proposal materials.
Recommended configuration patterns
I think it is easiest to start with these four categories.
Department-specific Notebooks
Development, Sales, Marketing, HR, etc.Project-specific Notebooks
Project A, Project B...Client-specific Notebooks
Client X, Client Y...Company-wide common Notebooks
Product overviews, internal rules, brand guidelines
You can start from here and subdivide them as needed while you use them.
NotebookLM is convenient because it is easy to share within an organization.
Since NotebookLM can be easily shared internally, let's actively proceed with creating a team Wikipedia.
Also, I recommend appointing an operations manager for each NotebookLM and updating the content regularly.
*Please also see this article regarding operational design for how to divide NotebookLM notebooks.
Finally, here is a summary.
If you use NotebookLM as an internal Wikipedia,
Do not group Notebooks too broadly; divide them into some detail.
Structure them along two axes: by department and by theme.
Keep the nature of the information in each Notebook consistent.
Just by following these three points, you will have an internal knowledge base that is highly accurate and easy to use in practice.
Thank you for reading!
Click here for past AI-related articles!
By the way, I have also published a book summarizing AI promotion in organizations. Please take a look if you like.
<Self-Introduction>
Until recently, I was at CARTA HOLDINGS, a group of over 20 operating companies with approximately 1,400 employees belonging to the Dentsu Group. There, I led AI utilization across the entire organization through the company-wide AI Promotion Office, while also serving as the Representative Director of D-Marketing Academy, a corporate training service for "Generative AI & Digital Marketing Talent," where I supported AI talent development for hundreds of companies ranging from large corporations to startups.
Currently, I serve as the Representative Director of the "AI Digital Community (ADC)," a community for AI promotion and digital-related companies, as well as Representative Director of FURIKAKE Partners Inc., which supports the "×AI" transformation of client businesses, and AI Portalize Inc., which provides products that support organizational AI utilization, supporting organizational AI utilization from various perspectives!
<Biography>
May 2005: Started an e-commerce business while in university.
May 2007: Joined CyberAgent, Inc. and was involved in launching new businesses.
October 2011: Established Flessel Inc. at VOYAGE GROUP, Inc. to conduct a collaborative business with KDDI, and assumed the position of Representative Director.
November 2015: Assumed the position of Representative Director of JS Consulting Inc., which conducts e-commerce consulting business.
April 2018: JS Consulting was acquired by Hamee Corp., a company listed on the Tokyo Stock Exchange Prime Market, and I continued as Representative Director.
May 2019: Appointed as Executive Officer of Hamee Corp., overseeing the new business domain of the Hamee Group.
February 2021: Appointed as Advisor to THE CHOSEN ONE Inc., which provides D2C support.
March 2021: Appointed as Director of NAAFY Inc., which conducts apparel D2C business.
April 2021: Established D-Marketing Academy Inc. and assumed the position of Representative Director.
January 2023: D-Marketing Academy was acquired by CARTA HOLDINGS, Inc., and I continued as Representative Director.
March 2025: Began concurrently serving in the AI Promotion Office, which promotes AI utilization across the entire CARTA HOLDINGS group.
January 2026: Established FURIKAKE Partners Inc., which provides advisory services regarding generative AI, and assumed the position of Representative Director.
January 2026: Established AI Portalize Inc., an organizational generative AI platform service, and assumed the position of Representative Director.
January 2026: Established the "AI Digital Community (ADC)," a community for digital-related AI utilization companies, and assumed the position of Representative Director.