NotebookLM in the Classroom: The Complete Methodology of “Creativity” Ltd.
NotebookLM in the Classroom: The Complete Methodology of “Creativity” Ltd. (Part 1)
We are pleased to launch a new, in-depth series of articles dedicated to NotebookLM – the tool that we at “Creativity” Ltd. (https://cpocreativity.com/) believe represents the safest, most effective, and pedagogically sound bridge between artificial intelligence and the real classroom.
Our expertise in this area is not theoretical. It is forged in practice. As pioneers in the implementation of AI in Bulgarian education, we have been testing this tool since its early beta versions. Our experience is part of our broader contribution to AI pedagogy, including through the development of university courses like “AI in the Primary Stage of Education”. But of all the tools we stress-tested in a real-world environment, NotebookLM stood out for its immediate applicability.
Why NotebookLM? The Vision We Share and Build Upon
When Google presented their vision for AI in education in the foundational document “AI and Learning: A new chapter for students and educators“, they formulated a philosophy that we at “Creativity” Ltd. not only share but actively develop. The purpose of AI is not to be a “quick answer” machine, but a catalyst that helps students “understand, comprehend, and apply” information.
NotebookLM is the exact embodiment of this philosophy. It is designed to help you “learn better by using your own sources”. We saw this as an opportunity to transform passive content into interactive tests, personalized flashcards, and immersive audio experiences, without leaving the safe framework of the curriculum.
Our Methodology: The Power of “Grounded AI”
At the core of our methodology is the concept of “grounded” AI. Unlike general chatbots, NotebookLM does not “hallucinate” because it does not have access to the entire internet while working with your files.
Lesson from the practice of “Creativity”:
In our early tests, we found that teachers often worry that AI might give “harmful advice” to students. With NotebookLM, this risk is eliminated. If the information is not in your PDF files or uploaded lessons, the model simply replies: “I cannot find this information in the provided sources.” This is a pedagogical victory – it teaches returning to the source, not making things up.
What does this mean for the Bulgarian teacher?
- Full Control Over Content: You are the curator. You decide whether the “knowledge base” will contain only one story by Yovkov or the entire history of Bulgaria for the 10th grade.
- Absolute Safety: Students work in a closed, controlled environment. There is no risk of encountering inappropriate external content.
- Focus on Analysis: Since the facts are given, the focus shifts to their interpretation and connection.
Beyond Text: Directing AI “Hosts”
One of the most innovative parts of our methodology is working with the multimedia capabilities of NotebookLM. We don’t just use the “Audio Overview” feature; we direct it.
Through hundreds of experiments, we found that if we assign the right role to the AI hosts, we can create engaging content in minutes. The tool can process transcripts from YouTube videos and turn them into dialogues.
Here is an example of our “recipe” (prompt) that turns dry text into an interesting podcast:
“Act as a host of an educational podcast for teenagers. Do a 5-minute Deep Dive on this historical document. One host should be skeptical and ask tough questions, and the other should respond with facts from the text, using modern analogies.”
This approach changes the game for students who learn better by listening (auditory learners) and is an indispensable assistant for resource support.
Proven in Practice: Real Results
Our confidence is not hypothetical. It is built on real applications in dozens of Bulgarian schools. On our website, we already share specific models that work today.
The flagship of our methodology is the model for preparation for the State Matriculation Exam (DZI) in Bulgarian Language and Literature (BEL). In this specialized topic (see the full analysis here), we demonstrate how NotebookLM becomes a tireless “matriculation assistant,” loaded with all the literary works, critical analyses, and evaluation criteria.
“After the ‘Creativity’ training for NotebookLM, my time for creating tests on new works was reduced from 2 hours to 15 minutes. But more importantly, the students started generating their own questions for self-assessment.”
– Shared by a teacher of Bulgarian Language and Literature (BEL), who completed our training.
This model is applicable everywhere:
- In kindergarten: Generating personalized fairy tales in which the children from the group are characters, based on the educational focus for the week.
- In the primary stage: A “reading assistant” that explains complex words from the reader in accessible language.
- At university: A tool for comparative analysis of multiple scientific articles for theses.
What’s next in the series?
This is just an introduction to our complete methodology. In the next articles, we will open our “kitchen” and share specific tools:
- Part 2: Specific examples and “recipes” for the primary and lower-secondary stage.
- Part 3: Deep technical analysis of our model for the DZI in BEL – source structure and specific prompts.
- Part 4: Our “golden” prompts for teachers – ready-made templates for your work.
