01
The AI tutor stack
1 · ChatGPT
Best for interactive tutoring, explanations, problem solving, practice and feedback. Use Study mode for guided, step-by-step learning instead of simply asking for answers.
2 · Gemini
Best for multimodal learning, visual explanations, quizzes and structured tutoring. Use Guided Learning or Study Notebooks to identify knowledge gaps and build personalized lessons.
3 · NotebookLM
Best for learning directly from your own materials. Upload notes, PDFs, articles, slides, websites and other sources, then generate study guides, quizzes, flashcards, Audio Overviews, Video Overviews and explanations grounded in those sources.
4 · Claude
Best for deeply analyzing long readings, explaining difficult concepts, discussing ideas and working through large documents. Where available, Learning mode focuses on reasoning rather than simply providing answers.
02
Define your goal
Tell AI exactly what you want to learn. Include:
- Your current level
- Topic
- Desired skill level
- Deadline
- Available study time
- Exam or real-world goal
- Materials you already have
Prompt:
I want to master [topic]. I am currently at [level]. My goal is [goal]. I can study [time] per day for [duration]. Act as my personal tutor.
03
Take a diagnostic test
Do not start studying blindly. Ask AI to test what you already know.
Prompt:
Give me a diagnostic assessment on [topic]. Ask one question at a time. Identify my strengths, weaknesses, misconceptions, and missing prerequisites.
04
Build your curriculum
Turn your weaknesses and goals into a structured learning path. Ask AI to divide the subject into:
- Fundamentals
- Core concepts
- Intermediate topics
- Advanced topics
- Practice
- Projects
- Revision
Prompt:
Based on my diagnostic results, create a personalized learning roadmap. Put prerequisite concepts first and progressively increase difficulty.
05
Learn one concept at a time
Avoid asking AI to explain an entire subject at once. For every concept, request:
- Simple definition
- Intuition
- Real-world analogy
- Detailed explanation
- Example
- Common mistakes
- Practice question
Prompt:
Teach me [concept] from first principles. Start simple, gradually increase the depth, and check my understanding before continuing.
06
Use the Socratic method
Do not let AI immediately reveal answers. Make it guide your reasoning.
Prompt:
Do not give me the answer immediately. Ask questions and give progressively stronger hints until I can solve it myself.
07
Use your own materials
Upload:
- Textbooks
- Lecture notes
- PDFs
- Research papers
- Slides
- Syllabi
- Assignments
- Practice tests
Use NotebookLM when you want learning heavily grounded in these sources. Prompt:
Teach me this material using only the provided sources. Identify the most important concepts and quiz me on them.
08
Ask for multiple explanations
If something does not click, change the explanation. Try:
- Explain like I am a beginner.
- Explain visually.
- Use an analogy.
- Give a practical example.
- Explain mathematically.
- Explain from first principles.
- Compare it with something I already know.
Never confuse one bad explanation with an inability to understand the concept.
09
Practice active recall
Reading explanations repeatedly can create an illusion of understanding. Close the notes and retrieve information from memory. Ask AI for:
- Short-answer questions
- Concept questions
- Flashcards
- Fill-in-the-blank questions
- Problem-solving exercises
- Mixed quizzes
Prompt:
Test my ability to recall and apply this topic without showing me the answers first.
10
Use adaptive quizzing
Make difficulty depend on your performance.
Prompt:
Quiz me one question at a time. Increase difficulty when I answer correctly. When I make a mistake, diagnose why and give me another question testing the same concept.
11
Build a mistake log
Your mistakes tell you what to study next. Ask AI to track:
- Question
- Your answer
- Correct answer
- Why you were wrong
- Underlying misconception
- Concept to review
- Follow-up question
Then regularly ask:
What patterns do you see in my mistakes?
12
Use the Feynman technique
Explain the topic back to AI in your own words.
Prompt:
I will explain [topic] as if I were teaching a beginner. Identify anything incorrect, incomplete, vague, or poorly understood.
If you cannot explain something simply, investigate that gap.
13
Practice application
Understanding is stronger when you can use knowledge in unfamiliar situations. Ask for:
- Case studies
- Real-world problems
- Simulations
- Projects
- Experiments
- Coding exercises
- Scenario questions
Prompt:
Give me a realistic problem where I must apply what I learned without telling me which concept to use.
14
Simulate an exam
Once you understand the material, remove the tutoring assistance.
Prompt:
Create a realistic closed-book practice exam on [topic]. Do not provide hints or answers until I submit the entire test.
Afterward, ask AI to analyze your weak areas.
15
Use spaced review
Do not review everything equally. Revisit difficult material more frequently and mastered concepts less frequently. Create review sessions around:
- Recent mistakes
- Weak concepts
- Older material
- Mixed retrieval
- Cumulative quizzes
Ask AI to generate a review plan based on what you struggled with.
16
Create a mastery loop
Repeat:
- Learn
- Recall
- Practice
- Get feedback
- Correct mistakes
- Explain it yourself
- Apply it
- Review later
Move forward when you can retrieve, explain, and apply the concept, not simply recognize it.
17
The 30-minute AI study session
- 5 minutes
- Review old material from memory.
- 10 minutes
- Learn one new concept.
- 10 minutes
- Solve problems without seeing answers.
- 5 minutes
- Review mistakes and summarize what you learned.
18
The 60-minute AI study session
- 10 minutes
- Active recall of previous material.
- 20 minutes
- Learn new concepts.
- 20 minutes
- Practice increasingly difficult problems.
- 5 minutes
- Correct mistakes.
- 5 minutes
- Create your next review plan.
19
Your master AI tutor prompt
Copy, paste and fill in the brackets:
Act as my personal tutor for [subject].
My current level is [level].
My goal is [goal].
My deadline is [date].
I can study [time] per day.
First, assess my current knowledge.
Then create a personalized curriculum based on my weaknesses.
Teach one concept at a time using simple explanations, analogies, examples, and progressively deeper explanations.
Use Socratic questions instead of immediately giving me answers.
After every concept, test me with active-recall and application questions.
Ask one question at a time and wait for my response.
When I make a mistake, explain why, identify the underlying misconception, and give me another problem testing the same skill.
Gradually increase the difficulty as I improve.
Regularly revisit concepts I struggle with.
Your goal is not to help me finish quickly. Your goal is to make sure I can understand, remember, explain, and apply what I learn.
20
Use the right AI for the right job
- ChatGPT
- Interactive tutoring, Study mode, reasoning practice, explanations, problem solving, quizzes and feedback.
- Gemini
- Guided Learning, Study Notebooks, visual learning, multimodal explanations, diagnostics, quizzes and adaptive lessons.
- NotebookLM
- Source-grounded learning, course materials, study guides, flashcards, quizzes, Audio Overviews, Video Overviews and research.
- Claude
- Long documents, deep explanations, discussion, writing feedback, conceptual analysis and reasoning-focused learning.
21
Ask better questions
Don’t use AI like this
“Give me the answer.”
Use AI like this
“Help me figure out the answer myself.”
Don’t just ask
“What is this?”
Ask
“Explain it, test me on it, find my misunderstanding, and make me apply it.”
22
Verify important information
AI systems can still make mistakes. For high-stakes or technical learning:
- Check primary sources
- Compare answers with textbooks or course materials
- Ask for sources
- Verify calculations independently
- Challenge suspicious claims
Never assume confidence means correctness.
23
The golden rule
Do not use AI to avoid thinking. Use AI to:
- Make yourself think harder
- Get faster feedback
- Expose knowledge gaps
- Practice more effectively
- Progressively become capable of solving problems without AI
The goal is mastery
The goal of an AI tutor is not dependence.
- Learn
- Recall
- Practice
- Feedback
- Correct
- Explain
- Apply
- Review
The goal is mastery: to understand, remember, explain and apply what you learn, and progressively solve problems without AI.
