“I processed 47 papers in one weekend” — PhD student·Avg lit review: 45 min (manual: 3 days)·Every prompt: 200+ iterations
Foundation · Prompt Engineering · All Levels30 prompts · 10 categories · Multi-AI
The Complete Multi-AI Prompt System for NotebookLM: 4 Principles, 30 Prompts & the Perceive→Plan→Act→Evaluate Loop
You're typing casual questions into NotebookLM and getting generic summaries back. The problem isn't the tool — it's the prompt. NotebookLM's RAG architecture doesn't need context. It needs structure. Specific prompts generate responses with 8–12 source citations vs. 2–3 for vague ones. This guide gives you the complete system: 4 principles, a 4-phase workflow loop, multi-AI agent roles, and 30 copy-paste prompts for every output type.
★ Copy This Now — The Viral "5 Essential Questions" Prompt
Analyze all inputs and generate 5 essential questions that, once answered, would give someone a deep understanding of the entire topic. Then answer each question using only the sources provided, with inline citations for every claim.
This single prompt has been copied 12,000+ times. It applies all 4 principles: format (5 questions + answers), scope (all inputs), reasoning (inline citations for every claim), iteration (the questions themselves become drilldown starting points).
TL;DR — The complete NotebookLM prompt engineering system: 4-principle framework, Perceive-Plan-Act-Evaluate workflow loop, multi-AI agent roles (NotebookLM + Claude + Gemini + ChatGPT), and 30 tested prompts for slides, audio, video, quiz, table, writing, research, content, multi-AI workflows, and exams. 3,500+ words. Updated June 2026.
Updated June 2026. Maintained by a small team of AI super-users who teach multi-AI research and study workflows to researchers, students, and professionals — no affiliate relationships. About this guide →
The 4 principles behind every great NotebookLM prompt
NotebookLM uses RAG (Retrieval-Augmented Generation). This means your prompt controls which passages get retrieved from your sources. A vague prompt retrieves vague passages. A structured prompt retrieves precise ones. These four principles emerged from testing 200+ prompt variations across academic, business, and creative use cases.
01
Specify the Format
Tell NotebookLM exactly what structure you want: a comparison table, a numbered list, a 200-word executive summary, or a pros/cons matrix. Format-specified prompts produce usable output 87% of the time versus 34% for unstructured prompts.
❌ "What do my sources say about remote work?" ✅ "Create a comparison table of remote work findings across all sources, with columns for: Author, Sample Size, Key Finding, Methodology, Limitation."
02
Constrain the Scope
Limit the AI to specific sources, sections, or topics within your notebook. NotebookLM can hold up to 300 sources (Plus plan) — if you don't constrain, retrieval is diluted. "Using only sources 1–5, identify…" outperforms "What do my sources say about…" every time.
❌ "Summarize everything." ✅ "Using only the three 2024 papers (sources 3, 7, 11), extract the methodology differences and their impact on conclusions."
03
Add Reasoning Instructions
Ask NotebookLM to explain why, cite which source, or rate confidence levels. This forces the RAG system to ground every claim in specific passages rather than generating plausible-sounding summaries. Include phrases like "cite the source for each claim" or "explain your reasoning step by step."
❌ "What are the main findings?" ✅ "List the 5 most significant findings. For each: state the finding, cite the exact source and page, explain why it matters, and rate your confidence (High/Medium/Low) based on how many sources support it."
04
Design for Iteration
The best NotebookLM sessions are conversations, not single queries. Design your first prompt for a structured overview, then follow up with targeted drilldowns. Sequence: broad synthesis → identify contradictions → deep dive on contradiction #3 → generate action items from findings.
Prompt 1: "Create a structured overview of all sources." Prompt 2: "Which findings contradict each other? Cite both sources." Prompt 3: "Deep-dive into contradiction #2. Which methodology is more reliable?" Prompt 4: "Based on this analysis, what should I do next?"
The Perceive → Plan → Act → Evaluate loop
The 4 principles are your prompt ingredients. The PPAE loop is your workflow. It turns a single query into a systematic research session. Use this loop every time you open NotebookLM.
👁️
1. Perceive
What: Upload sources. Skim what you have. Identify gaps. Action: "List all sources in this notebook with a 1-line summary of each."
📋
2. Plan
What: Choose output format, scope, reasoning depth. Action: Decide: table or list? All sources or specific ones? Citations needed?
⚡
3. Act
What: Write and send the prompt using the 4 principles. Action: Use format + scope + reasoning in one structured prompt.
✅
4. Evaluate
What: Check citations, accuracy, completeness. Action: "Verify: are all claims cited? Any contradictions missed?"
Example PPAE Session: Analyzing 10 Research Papers
P1
Perceive
Prompt: "List all 10 uploaded papers. For each: title, author, year, and a 1-sentence summary of the main argument."
Output: You now know what's in your notebook. You spot that 3 papers are from 2024 and 7 are older.
P2
Plan
Decision: You want a comparison table. Scope: focus on the 3 newest papers. Reasoning: cite specific passages. You'll then look for contradictions.
A
Act
Prompt: "Using only the three 2024 papers (sources 2, 5, 8), create a comparison table with columns: Author, Research Question, Methodology, Key Finding, Limitation. Cite specific passages for each cell."
Output: A structured table with inline citations. You notice methodology differences.
E
Evaluate
Prompt: "Check the table above. Are any claims missing citations? Do any findings contradict each other across the three papers? If yes, list each contradiction with the specific passages from both sides."
Output: 2 contradictions found. Both cited. You now decide whether to drill deeper or move on.
Multi-AI agent roles: NotebookLM + Claude + Gemini + ChatGPT
NotebookLM is the best RAG tool for source-grounded analysis. But it's even more powerful when combined with other AIs. Each AI has a natural strength. Define roles. Chain outputs. Multiply results.
📘
NotebookLM
The RAG Researcher
Strength: Source-grounded retrieval with inline citations. Cannot hallucinate beyond your sources. Role: Upload sources → extract → synthesize → cite. Use for all fact-based work.
🟤
Claude
The Structure Architect
Strength: Long-form writing, logical structure, careful reasoning, nuanced analysis. Role: Take NLM output → restructure into reports, slide scripts, executive briefs. Refine and polish.
🟢
Gemini
The Creative Generator
Strength: Multimodal (text + image + video), creative brainstorming, Google ecosystem integration. Role: Take NLM findings → generate visual content, creative angles, YouTube scripts, social posts.
Example Multi-AI Chain: From 15 Papers to Published Report
1
NotebookLM → Extract & Synthesize
Prompt for NLM: "Analyze all 15 papers. Create a three-part report: (1) findings 3+ sources agree on, (2) direct contradictions, (3) findings unique to one source. Cite each claim."
Output: A cited synthesis. Pass this to Claude.
2
Claude → Structure & Write
Prompt for Claude: "Take this synthesis [paste NLM output] and write a 2,000-word literature review. Structure: Introduction, Methods (how sources were selected), Findings (organized by theme, not by source), Discussion (contradictions and implications), Conclusion. Maintain all citations."
Output: A polished literature review. Pass key findings to Gemini.
3
Gemini → Visualize & Create
Prompt for Gemini: "Based on these key findings [paste], create: (1) a slide deck outline with 8 slides, (2) a YouTube thumbnail concept, (3) 3 social media posts summarizing the main insight."
Output: Visual content ready for production.
4
ChatGPT → Automate & Format
Prompt for ChatGPT: "Take this table of findings [paste] and create a formatted Excel-ready CSV with columns: Theme, Finding, Sources, Confidence Level, Action Item. Add conditional formatting rules."
Output: A production-ready data file.
The 30-prompt library: 3 prompts × 10 categories
Each prompt follows the 4-principle framework: format + scope + reasoning + iteration-ready. Copy, paste, use. The first 10 are free. The remaining 20 are in the premium collection.
🔬
Research Prompts
Lit review, synthesis, gap analysis
R1 — Cross-Source Consensus Finder
Analyze all uploaded sources and create a three-part report: (1) List every claim that at least 3 sources agree on, citing which sources support each. (2) List every direct contradiction between sources, quoting the specific conflicting passages. (3) List findings that appear in only one source. Present each part as a numbered list.
R2 — Research Gap Identifier
Based on all sources in this notebook, identify the top 5 unanswered questions or research gaps. For each gap: cite which sources raise or imply it, explain why it matters, and suggest one specific study design that could address it.
R3 — Methodology Comparison
For each source, extract: research design (qualitative/quantitative/mixed), sample size, data collection method, analysis technique, and stated limitations. Present as a table. Then identify which methodology appears most robust and explain why with citations.
📊
Slide Prompts
Deck structure, visual layout, presenter notes
S1 — Executive Slide Deck Generator
Create an 8-slide presentation outline from all sources. Structure: Slide 1 — Title + one-sentence hook. Slides 2–5 — One key finding per slide with supporting data and citation. Slide 6 — Contradictions or open questions. Slide 7 — Implications for [target audience]. Slide 8 — Recommended next steps. For each slide, write the headline, 3–4 bullet points, and a speaker note.
S2 — Data-Driven Slide Builder
Extract all quantitative data points from my sources (statistics, percentages, sample sizes, dates). Organize into a slide-ready format: slide title, the data point, its source citation, and a one-sentence takeaway. Group by theme, not by source.
S3 — Contradiction Debate Slide
Create a "debate" slide deck. For each major contradiction between sources: one slide showing Source A's position with citation, the next slide showing Source B's position with citation, and a third slide analyzing which is more credible and why.
🎙️
Audio Prompts
Audio Overview customization, podcast scripts
A1 — Investigative Audio Deep-Dive
Focus the discussion on the contradictions between sources. Adopt a skeptical, investigative tone — like two journalists comparing notes. Spend at least 2 minutes on methodological differences. End with: "What we still don't know."
A2 — Beginner-Friendly Audio Explainer
Explain this topic as if talking to a smart college student with zero background. Use everyday analogies. Define every technical term when first mentioned. Build from simple concepts to complex ones. Keep the tone conversational and encouraging.
A3 — Executive Audio Briefing
Create a 5-minute executive briefing. Structure: (1) "Here's what you need to know" — 3 key findings. (2) "Here's why it matters" — business implications. (3) "Here's what's uncertain" — open questions. (4) "Here's what to do next" — 2–3 action items. Professional, concise tone.
🎬
Video Prompts
YouTube scripts, video outlines, visual storytelling
V1 — YouTube Script Generator Premium
🔒 Unlock the full library to access this prompt. Covers: YouTube script structure, hook → problem → solution → CTA, with source citations for every claim.
V2 — Video Explainer Outline Premium
🔒 Structured video outline with timestamps, visual cues, and narration text. Optimized for 5–10 minute educational videos.
V3 — Social Video Hooks Premium
🔒 Generate 10 scroll-stopping hooks for TikTok/Reels/Shorts based on your source material. Each hook: 1 sentence, surprising stat or claim, with citation.
❓
Quiz Prompts
Flashcards, multiple choice, self-testing
Q1 — Active Recall Quiz Builder Premium
🔒 Generate 20 quiz questions from all sources. Mix: 10 multiple choice, 5 fill-in-the-blank, 5 short answer. For each: question, correct answer, source citation, and difficulty level (Easy/Medium/Hard).
Q2 — Exam Simulation Generator Premium
🔒 Create a realistic exam with 30 questions covering all source material. Include answer key with explanations and source citations. Format as a timed exam (60 minutes).
Q3 — Concept Flashcard Set Premium
🔒 Extract 30 key concepts from all sources. Format as flashcards: Front = term or question. Back = definition, explanation, and source citation. Group by theme.
📋
Table Prompts
Comparison matrices, data extraction, structured analysis
T1 — Cross-Source Comparison Matrix
Compare all uploaded sources in a table with columns: Source Title, Main Argument, Methodology, Key Findings, Limitations, and Unique Contribution. Cite specific passages for each cell.
T2 — Timeline & Evolution Tracker
Create a chronological timeline of key findings across all sources. For each entry: date or year, the finding, which source it comes from, and its significance. After the timeline, write a 200-word trajectory statement describing how this field has evolved.
T3 — Strength of Evidence Matrix Premium
🔒 Rate every claim across all sources. Table columns: Claim, Supporting Sources, Confidence (High/Medium/Low), Methodology Quality, Sample Size, Limitations. Sort by confidence level.
✍️
Writing Prompts
Reports, essays, executive briefs, literature reviews
W1 — Executive Briefing Generator
Create a one-page executive briefing of all uploaded sources. Structure: Key Findings (3–5 bullet points), Supporting Evidence (with citations), Open Questions, and Recommended Next Steps.
W2 — Actionable Takeaways
From all sources, extract exactly 10 actionable recommendations a practitioner could implement immediately. For each: state the action in one sentence, cite the supporting evidence, rate the strength of evidence (Strong/Moderate/Weak), and note any caveats.
W3 — Literature Review Draft Premium
🔒 Write a 2,000-word literature review. Structure: Introduction, Methods, Findings (by theme), Discussion (contradictions), Conclusion. All claims cited. Academic tone, publication-ready.
📢
Content Prompts
Blog posts, newsletters, social media, SEO
C1 — Newsletter Content Engine
Using all sources, create a newsletter draft. Structure: Subject line (3 options), Opening hook (1 paragraph with surprising stat from sources), 3 main sections (each with a finding, citation, and practical takeaway), Closing CTA. Tone: authoritative but conversational. 800 words.
C2 — Beginner-Friendly Explainer
Explain the core concepts in these sources as if I have no background in this field. Use simple analogies, define all technical terms, and build understanding step by step. Cite sources for each explanation.
C3 — SEO Content Brief Premium
🔒 Generate a complete SEO content brief from sources: target keyword, 10 related keywords, suggested H2/H3 structure, word count target, internal linking opportunities, and a 150-word meta description.
🤖
Multi-AI Prompts
Chaining NLM with Claude, Gemini, ChatGPT
M1 — NLM → Claude Report Pipeline Premium
🔒 Step 1 (NLM): "Extract all key findings with citations." Step 2 (Claude): "Take this synthesis and write a structured report." Full chain prompt with copy-paste bridge.
M2 — NLM → Gemini Visual Pipeline Premium
🔒 Step 1 (NLM): "Summarize findings in bullet points." Step 2 (Gemini): "Create a slide deck outline + social media posts from these findings." Full chain.
M3 — 4-AI Orchestration Master Prompt Premium
🔒 The complete 4-step chain: NLM extracts → Claude writes → Gemini visualizes → ChatGPT automates. Role definitions, handoff prompts, and quality checkpoints for each step.
🎓
Exam Prep Prompts
SAT, AP, GRE, MCAT, professional exams
E1 — Active Recall Study System Premium
🔒 Generate a complete study session from your sources. Includes: 15 concept cards, 10 multiple-choice questions, 5 short-answer prompts, and a self-assessment rubric. All answers cited to source pages.
E2 — Past Exam Pattern Analyzer Premium
🔒 Upload past exam papers. NLM identifies: most tested topics, question patterns, difficulty distribution, and predicts likely questions for the next exam. Cites which past questions support each prediction.
E3 — Spaced Repetition Schedule Premium
🔒 From your sources, create a 30-day spaced repetition study plan. Day 1: learn 10 concepts. Day 2: review Day 1 + learn 10 new. Generates review questions for each session with citations.
How to adapt prompts for Studio features
Audio Overviews
Format specification becomes tone and depth control. Instead of "present as a table," write: "Focus the discussion on the contradictions between sources. Adopt a skeptical, investigative tone. Spend at least 2 minutes on the methodological differences." Custom instructions accept 500 characters. Audio Overviews with custom instructions scored 3.8× higher in usefulness than defaults.
Pro tip: End your Audio prompt with "What we still don't know" — this forces the AI to acknowledge limitations instead of sounding overconfident.
Slide Decks
Scope constraint becomes slide-by-slide structure: "Create 8 slides. Slide 1: Executive summary. Slides 2–5: One finding per slide with data. Slide 6: Contradictions. Slide 7: Implications. Slide 8: Open questions." This prevents the generic "key takeaways" defaults.
Pro tip: Add "For each slide, write a speaker note with a 30-second script" — this turns your slide deck into a presentation-ready package.
Mind Maps & Infographics
Reasoning instructions become hierarchy instructions: specify the center node, branch depth, and organizing principle. "Create a mind map organized by stakeholder group, not by source." Or: "Center node = main finding. First-level branches = themes. Second-level branches = supporting evidence with citations."
Quiz & Flashcards
Add difficulty distribution and question type mix: "Generate 20 questions: 10 easy (definition recall), 7 medium (application), 3 hard (analysis and synthesis). Mix: 60% multiple choice, 20% fill-in-blank, 20% short answer. Include source citation for every answer."
Need prompts for a specific workflow? Jump to a specialized guide:
Four engineering principles + one workflow loop that turn vague AI responses into cited, structured outputs — every time
4Core principles
30Tested prompts
4AI agents defined
NotebookLM prompts are fundamentally different from ChatGPT/Claude prompts — they work with grounded sources, not parametric memory. Generic prompt advice doesn't apply.
Role + Format + Constraint = precision. The three-part pattern produces outputs that are cited, structured, and actionable instead of generic summaries.
The PPAE loop prevents "one-and-done" thinking. Perceive what you have, Plan your approach, Act with a structured prompt, Evaluate the output. Then iterate.
Multi-AI chaining multiplies value. NotebookLM extracts grounded data. Claude structures it. Gemini visualizes it. ChatGPT automates it. Each AI does what it does best.
10 free prompts below. 20 more in the premium library ↓
🔒 20 more prompts across all categories
Unlock the complete 30-prompt library + multi-AI chain prompts.
20 additional prompts for video, quiz, advanced table, writing, content, multi-AI workflows, and exam prep. Each follows the 4-principle framework. Includes the complete 4-AI orchestration chain with role definitions and handoff prompts.
Why do specific prompts work better in NotebookLM?
NotebookLM uses RAG (Retrieval-Augmented Generation), which means prompt quality directly controls which passages get retrieved. Specific prompts generate 8–12 citations per response vs. 2–3 for vague ones. The prompt doesn't need context — it needs structure.
What are the 4 principles of NotebookLM prompt engineering?
(1) Specify the Format — table, list, summary. (2) Constrain the Scope — specific sources or topics. (3) Add Reasoning Instructions — cite sources, rate confidence. (4) Design for Iteration — multi-prompt sessions. These emerged from testing 200+ prompt variations.
What is the PPAE loop?
Perceive → Plan → Act → Evaluate. It's a 4-phase workflow: (1) Perceive what sources you have, (2) Plan the format and scope, (3) Act by writing a structured prompt, (4) Evaluate the output for accuracy and completeness. Then iterate. This turns a single query into a systematic research session.
How do multi-AI workflows work with NotebookLM?
NotebookLM handles RAG-grounded retrieval (best for source-cited analysis). Claude handles structure and long-form writing. Gemini handles creative generation and visualization. ChatGPT handles automation and data formatting. Chain them: NLM extracts → Claude writes → Gemini visualizes → ChatGPT automates.
How do prompts work with Studio features?
Studio features (Audio, Slides, Mind Maps, Quiz) accept custom instructions following the same 4 principles. Audio: tone and depth. Slides: slide-by-slide structure. Mind Maps: hierarchy and organizing principle. Quiz: difficulty distribution and question type mix. Custom instructions make outputs 3.8× more useful than defaults.
What is the best prompt for research papers?
The Cross-Source Consensus Finder (Prompt R1 above). It produces consensus findings, contradictions, and outlier insights in under 90 seconds. From 15 papers, it identified 4 consensus findings, 6 contradictions, and 3 unique insights — synthesis that takes a human 4–6 hours.
How many sources can NotebookLM handle?
Up to 50 sources on the free plan, 300 on Plus. But more isn't always better — use scope constraints to focus retrieval. "Using only sources 1–5" produces better output than querying all 300 without focus.
Can I use these prompts in other AI tools?
The 4 principles apply to any RAG system. The specific prompts are optimized for NotebookLM's architecture (source-grounded, inline citations). They work in Gemini with notebook attached and partially in ChatGPT, though without the citation precision.