🛠 The Toolkit
To make this work, we use a powerful combination of Google’s ecosystem:
- Google Gemini: The core engine that handles our complex queries and data analysis.
- Gems (Custom Personalities): Where we build the "tutor" persona, setting rules for how the AI should teach, quiz, and support.
- NotebookLM: A specialised environment where the AI is "grounded" strictly in the documents we provide, ensuring 100% relevance to our notes.
📋 The 4-Phase Process
In the attached document and video, I walk you through the four essential phases of setting this up:
- Phase 1: Knowledge Curation – Identifying your "Primary Truth" (PDFs, notes, or datasets).
- Phase 2: Persona Setup – Using Gems to define the tutor’s role and operational constraints.
- Phase 3: The Pedagogical Layer – Implementing Socratic guardrails so the AI guides you rather than just giving answers.
- Phase 4: Deployment – Linking everything into a seamless study interface.
🎥 Watch the Walkthrough
I’ve attached a video showing the exact steps I took to build a tutor for L&D platform analysis, along with a short PDF guide you can follow.
Question for the community: If you could "ground" an AI tutor in any single book or course you’ve taken, which one would it be and why? Let’s discuss how this could change your study habits below! 👇