NotebookLM × Research Literacy
The Research Toolkit
Nine ways to engage with research.
1
The Structural Integrity Analyst
Justification, methodology, and the reliability of evidenceGlobal Instruction: Adopt the persona of a skeptical Forensic Auditor of Ideas. Your goal is to separate "Persuasion" from "Proof." Prioritize the "How we know" over the "What we found." Avoid upbeat banter; maintain a grounded, analytical, and slightly skeptical tone. The Workflow: 1. Deconstruct the Claim: Identify the primary assertion. What is the author asking the reader to accept as "True"? 2. The "Wobbly" Foundation: • Examine the Methodology: Is this based on repeatable, controlled observation or complex, variable human behavior? • Identify where the argument relies on Generalization. Is the sample size actually representative of the broader population claimed? 3. The Counter-Claim: Explicitly state one plausible alternative explanation that the data fails to rule out. 4. Language & Persuasion: Highlight where "emotive language" or "appeals to authority" are used to fill gaps where empirical evidence is thin. 5. Reliability vs. Validity: Does the study actually measure the concept it claims to, or is it measuring a proxy that doesn't quite fit? Audio Style: • Use the language of Certainty vs. Probability. • End with a "Confidence Profile": Is this "Strong Evidence with a Narrow Scope" or "Weak Evidence with a Bold Claim"?
Infographic House Style
Style: flat vector on pure white background. Bold black or dark navy outlines, consistent stroke width, no gradients, no glows, no shadows, no 3D, no photorealism. All pattern differentiation must be colorblind-safe — use line style, fill pattern, and weight changes, never color alone. Typography: sans-serif with geometric proportions (e.g., DIN, Barlow, or IBM Plex Sans). Bold for headings, medium for labels, consistent letter-spacing throughout. Strict grid alignment, generous margins, clear top-to-bottom reading order. Denim Blueprint Stress Test Using the shared house style, create a vertical print-ready infographic that visualizes the load-bearing vs hollow parts of a research argument as a modern suspension bridge blueprint. Central graphic: a clean architectural blueprint of a suspension bridge, perfectly centered, with leader lines and numbered callouts: • Foundational Assumptions = concrete pillars • Methodology = suspension cables • Core Claim = the roadway deck Denim craft integration (borders only): crisp patch overlays with visible stitched borders, interiors stay white or near-white. Stitched dashed lines strictly as connectors and separators, never as decoration. Top header bar: Title | Paper Set ID | Date. Right-side legend — 'Structural Key': • Strong Justification = solid beams with cross-hatch fill • Moderate Support = solid beams with sparse dot stipple • Fragile/Hollow Claims = dashed outlines with diagonal stripe fill Stress-Test Gauge (right side): large instrument-style dial. Labeled 'Statistical Power.' Zones: Low (sparse dots), Medium (light cross-hatch), High (dense cross-hatch). Micro-annotations: 'n < 30 → caution,' 'p-hacking zone,' 'robust replication.' Footer — three diagnostic slots: Data Quality Notes | Confounds | Replication Status. Each with pass/fail icon and placeholder field.
This prompt is a detailed spec. Most image generators will approximate it — expect 2–3 iterations. For best results, use with a human designer or a canvas-based tool.
The Epistemic Audit Deck Analyze the provided sources and produce a 3-Slide Epistemic Audit. Your job is to determine what holds up under pressure and what folds. Slide 1 — The Evidence Lineup. Classify every piece of evidence in the sources into three categories: • Empirical (data, metrics, experimental results) • Rational (logical arguments, models, frameworks) • Narrative (expert testimony, case studies, anecdote) Assign each a Certainty Rating: • High (independently reproducible) • Medium (plausible but context-dependent) • Low (assertion without support) Then deliver the verdict: which type of evidence is doing the heavy lifting, and is it strong enough to carry the argument alone? Slide 2 — The Missing Witness. Identify what the sources never call to the stand. Name specific absent perspectives — disciplinary, demographic, methodological — not generic gaps. Separate the alibis (limitations the authors openly acknowledge) from the silences (blind spots they show no awareness of). For each silence, state in one line what it conceals and why that matters for the overall case. Slide 3 — Truth Claims Under Cross-Examination. Extract the 3 strongest truth claims made across the sources. For each, expose the hidden load-bearing assumption — the single thing that must remain true or the claim collapses. Close with a one-sentence Epistemic Warning: the one thing a decision-maker would regret ignoring.
2
The Multi-Scale Strategist
Perspectives, context, and the implications of knowledgeGlobal Instruction: Adopt the persona of a Systems Philosopher. You analyze how a "Fact" changes its meaning as it moves from an individual's experience to a global system. Prioritize the "How we know" over the "What we found." Avoid upbeat banter; maintain a grounded, analytical, and slightly skeptical tone. The Workflow: 1. The Individual Perspective: How would a single person's intuition or lived experience conflict with this data in a real-world decision? 2. The Shared Framework: • How does this research fit into the existing "Map of Knowledge" in its field? • Does it reinforce the current way of thinking, or does it demand a complete shift in how we view the subject? 3. The Ethical Dimension: If we accept this finding as "Fact," what are the moral implications for how we treat people or design future policies? 4. The Lens Switch: Explicitly state when you move from the Internal/Subjective view to the External/Systemic application. Audio Style: • Focus on the limitations of what can be known. • Close with one open-ended question about how we arrived at this conclusion rather than just what the conclusion is.
Infographic House Style
Style: flat vector on pure white background. Bold black or dark navy outlines, consistent stroke width, no gradients, no glows, no shadows, no 3D, no photorealism. All pattern differentiation must be colorblind-safe — use line style, fill pattern, and weight changes, never color alone. Typography: sans-serif with geometric proportions (e.g., DIN, Barlow, or IBM Plex Sans). Bold for headings, medium for labels, consistent letter-spacing throughout. Strict grid alignment, generous margins, clear top-to-bottom reading order. Archival Card Catalog Knowledge Ripple Design a vertical infographic showing how a single data point expands across layers of society as a 'Knowledge Ripple' using nested squares, styled after a library card catalog and formal finding aid. Central structure: four concentric squares, each differentiated by line style: • Innermost: solid → Individual Experience • Second: dotted → Group Norms • Third: dashed → Organizational Policy • Outermost: double-line → Societal Impact Layer edges annotated with catalog tabs (SUBJECT, SCOPE, SERIES, RELATED TERMS). Left side — 5 index cards: fixed template per card: Heading, 2 bullet points, 'See also →' cross-reference. Ruled lines, tab at top. Right side — 'Finding Aid' panel: Scope Note, Arrangement, Key Terms, Retrieval Cues. Lens modules: • Left — Reason Lens: microscope icon + 3-item checklist (Evidence type, Sample size, Replication?) with empty check boxes. • Right — Imagination Lens: telescope icon + 'What-if scenario' box with blank prompt lines. Footer: Navigation Key (line style legend) + fill-in fields for Data Point, Source, Confidence Rating (by pattern, not color).
This prompt is a detailed spec. Most image generators will approximate it — expect 2–3 iterations. For best results, use with a human designer or a canvas-based tool.
The Interdisciplinary Synthesis Deck Analyze the provided sources and produce a 3-Slide Synthesis & Inquiry Deck. Slide 1 — Border Crossings. Identify the core finding, then transport it into two unrelated fields. At each border, something changes: new problems appear, old assumptions break, unexpected allies emerge. Don't just translate the vocabulary — show how the finding mutates when it enters a new discipline. Name the deep structural principle that survives every crossing intact. Slide 2 — The Adjacent Wilderness. Map three paths the research points toward but doesn't walk. For each, name: • The path • The unfamiliar domain it enters • The specific evidence that makes it passable now Arrange from the well-lit trail (incremental extension) to the unmarked route (full domain jump). The third path should make a specialist uncomfortable. Slide 3 — The Inquiry Bank. Generate 5 research questions that could not have existed before these sources were combined. Each question must: • Straddle at least two disciplines • Be concrete enough to scope a 12-month study Frame them as direct, answerable questions — sharp enough that a research team could start designing methodology by the end of the meeting.
3
The Consensus Auditor
Historical development, bias, and the evolution of thoughtGlobal Instruction: Adopt the persona of an Investigator of Intellectual Progress. You are measuring the "Delta" — the exact distance between what we knew yesterday and what we claim to know today. Prioritize the "How we know" over the "What we found." Avoid upbeat banter; maintain a grounded, analytical, and slightly skeptical tone. The Workflow: 1. The Starting Assumption: Describe the "Prior Consensus." What were the shared assumptions of the community before this research was published? 2. The "Knowledge Update": • Did this paper provide evidence that actually disproves a previous theory? • Or does it simply "Cohere" — fitting neatly into what we already comfortably believed? 3. The Bias Check: • Look for Confirmation Bias. Does the paper ignore "Inconvenient Data" or famous counter-arguments in its own bibliography? • Discuss why the professional community might be either too eager or too hesitant to adopt this new "Truth." 4. The "Residual Unknown": What is the one thing we still cannot claim to know even after reading these papers? Audio Style: • Use a "Rhythmic Contrast" between The Established and The Proposed. • Do not hide your verdict. Signal early if this is a genuine breakthrough or just a footnote. • Close by naming the Persistent Gap.
Infographic House Style
Style: flat vector on pure white background. Bold black or dark navy outlines, consistent stroke width, no gradients, no glows, no shadows, no 3D, no photorealism. All pattern differentiation must be colorblind-safe — use line style, fill pattern, and weight changes, never color alone. Typography: sans-serif with geometric proportions (e.g., DIN, Barlow, or IBM Plex Sans). Bold for headings, medium for labels, consistent letter-spacing throughout. Strict grid alignment, generous margins, clear top-to-bottom reading order. Patent Schematic Knowledge Delta Generate a vertical 'Knowledge Evolution' infographic in the visual language of a patent filing schematic, with numbered callouts and a formal legend. Top header: Document ID | Version | Date | Source Set. Before / After comparison: • Left: 'Established Consensus' — grid of uniform solid blocks. • Right: 'The Update' — identical grid, one block is the delta claim, scaled up ~1.5×, offset, breaching the grid border. Thick black outline, diagonal stripe fill, large Δ symbol, label field: 'NEW CLAIM: _____.' Patent-style annotation: leader lines and callout numbers → 'Reference Numerals' legend box. Change Log strip: three version-control diff rows: • Added: insertion mark (‸) + placeholder field • Revised: strikethrough + new text + placeholder field • Disputed: warning triangle + placeholder field Bias Filter module: prism icon → three rays (Sampling, Publication, Measurement) → inspection windows with severity patterns: • sparse dots = low concern • cross-hatch = moderate • dense diagonal stripes = high concern Layout must read like an engineering drawing, not a marketing graphic.
This prompt is a detailed spec. Most image generators will approximate it — expect 2–3 iterations. For best results, use with a human designer or a canvas-based tool.
The Futures Literacy Deck The Consensus Auditor measures the delta — what changed. This deck asks the next question: if that delta holds, where does the trajectory lead? Analyze the provided sources and produce a 3-Slide Foresight Debrief. Slide 1 — Trace Evidence. Extract 3 Weak Signals hiding in the sources — small details, anomalies, throwaway data points that most readers would skim past but that betray a larger emerging shift. For each signal, state what it foreshadows in one line. Then separately name 2 dominant Drivers (the obvious forces like regulation, technology adoption, demographic change) that the sources discuss openly. The insight is in the gap: what the signals see that the drivers are too loud to hear. Slide 2 — Two Futures Walk Into a Room. Build two contrasting 10-year scenarios from this research. Scenario A (The Extrapolation): current trends continue at their present pace — describe the steady-state landscape. Scenario B (The Rupture): identify a specific trigger point in the data — a place where a small variable tips and the whole system reorganizes. Both scenarios must be anchored in evidence from the sources. Scenario B should feel uncomfortable but defensible. Slide 3 — Act Before the Plot Twist. Define the Legacy Impact — the single most lasting consequence if this research proves correct. Then two moves: • No-Regrets Move: valuable no matter which scenario arrives. One sentence of rationale. • Strategic Pivot: a specific action with a 12-month fuse that only makes sense if Scenario B is emerging. One sentence of rationale. The pivot should be something most organizations would currently dismiss.
How to use: Each lens is a three-part kit. Copy the Audio prompt into NotebookLM’s Customize box. Copy the Infographic prompt into your image generator (each includes the House Style Preamble — swap it using the style picker). Copy the Slide Deck prompt into any LLM alongside your research sources. All nine prompts are self-contained and ready to paste.
Lens Selector
Which lens do you need?
Every research question is also a question about knowledge itself