Copy Process these 50 documents as a single context: 1. Create master index of all topics 2. Cross-reference findings across documents 3. Identify contradictions and patterns 4. Generate unified summary with document citations Use full context window - don't truncate or summarize prematurely.

Copy Analyze this multimodal dataset: - PDF report [attached] - Product images [5 images] - Customer video reviews [3 videos] - Audio feedback [2 recordings] Cross-reference insights across ALL modalities. Identify patterns only visible when combining data types.

Copy Analyze this video with temporal markers: For each scene change: - Timestamp - Visual elements present - Spoken content - Body language/emotions - Key takeaways Create timeline showing evolution of concepts throughout video.

Copy Analyze spatial relationships in this image: 1. Object positions (relative coordinates) 2. Distance relationships between elements 3. Perspective and depth cues 4. Spatial patterns and arrangements 5. Architectural or layout implications Provide coordinate-based descriptions where relevant.

Copy Compare these 10 product images: Create comparison matrix showing: | Feature | Image1 | Image2 | ... | Image10 | |---------|--------|--------|-----|---------| Identify: - Common patterns across all images - Unique features in each - Quality differences - Optimal image for each use case Rank images by effectiveness for [specific purpose].

Copy REASONING MODE: ON Before answering, explicitly work through: 1. Problem decomposition 2. Step-by-step logic 3. Alternative approaches considered 4. Verification of reasoning Show your thinking process, then provide final answer.

Copy DEEP RESEARCH MODE: Research topic: [X] Requirements: - Explore 20+ diverse sources - Identify expert opinions and consensus - Map relationships between concepts - Surface contradictions in literature - Provide synthesis with source attribution Take time to be thorough - prioritize depth over speed.

Copy Solve this problem using verified chain-of-thought: STEP 1: [Reasoning] VERIFICATION 1: [Check logic] STEP 2: [Reasoning] VERIFICATION 2: [Check logic] Continue pattern... FINAL ANSWER: [Result] FINAL VERIFICATION: [Confirm entire solution]

Copy Generate 5 competing hypotheses for this problem: H1: [Hypothesis] → Test → Result → Likelihood H2: [Hypothesis] → Test → Result → Likelihood H3: [Hypothesis] → Test → Result → Likelihood H4: [Hypothesis] → Test → Result → Likelihood H5: [Hypothesis] → Test → Result → Likelihood Rank hypotheses by evidence strength. Recommend best hypothesis with confidence level.

Copy After generating response, run error detection: CHECK 1: Logical consistency CHECK 2: Factual accuracy CHECK 3: Mathematical correctness CHECK 4: Source reliability CHECK 5: Assumption validity Report: [Passed/Failed] for each check Fix any failures before finalizing response.

Copy Partition context into zones: ZONE 1 (Tokens 1-250K): Background/reference material ZONE 2 (Tokens 250K-500K): Supporting evidence ZONE 3 (Tokens 500K-750K): Detailed analysis ZONE 4 (Tokens 750K-1M): Current focus area When answering, specify which zones are relevant.

Copy Build knowledge graph from all documents: NODES: Key concepts from each document EDGES: Relationships between concepts ATTRIBUTES: Source document, page number, confidence When answering, trace reasoning path through graph. Cite: [Document X, Section Y] → [Document Z, Section W]

Copy Organize context hierarchically: LEVEL 1: Executive summaries (refer first) LEVEL 2: Detailed sections (drill down as needed) LEVEL 3: Supporting data (reference for specifics) LEVEL 4: Raw documents (verify details) Always start at LEVEL 1, descend only when necessary.

Copy Tag all information with temporal metadata: [CURRENT - 2025]: Latest data [RECENT - 2024]: Still relevant [DATED - 2023]: May be outdated [HISTORICAL - <2023]: Background only Prioritize CURRENT data. Flag when relying on DATED information.

Copy For multi-turn conversations: After each response, identify: - KEEP: Still relevant information - ARCHIVE: May be needed later - DISCARD: No longer relevant Actively manage context to maintain focus. Retrieve ARCHIVED info only when explicitly needed.

Copy For current information, use native search: Search for: [query] from reliable sources Synthesize findings with web citations Verify across multiple sources Flag if information conflicts Always prefer Google Search for events after June 2024.

Copy Access my Google Workspace data: - Review emails from [date range] about [topic] - Check calendar for meeting conflicts - Analyze Drive documents matching [criteria] - Draft response referencing specific conversations Provide summary with links to original sources.

Copy Analyze YouTube video: [URL] Extract: - Key points from transcript - Visual demonstrations shown - Timestamp for each major concept - Viewer questions from comments Synthesize into structured summary with time markers.

Copy Use Google Maps context: - Find locations matching [criteria] - Analyze foot traffic patterns - Compare neighborhoods by [metrics] - Route optimization considering real-time conditions Provide map links and location recommendations.

Copy Edit this Google Doc: [link] Mode: SUGGESTING (track all changes) 1. Fix grammatical errors 2. Improve clarity and flow 3. Add citations where needed 4. Suggest restructuring if needed Provide summary of all changes made with rationale.

Copy AUTONOMOUS TOOL MODE: Available tools: Search, Calculator, Code Execution, File Analysis Task: [complex multi-step problem] Decide which tools to use, in what order, without asking. Show tool usage log after completion.

Copy Create execution plan, then execute: PLAN: Step 1: [action] → Expected output Step 2: [action] → Expected output ... Execute plan step-by-step. Adapt if unexpected results occur. Report progress after each step.

Copy Write and execute code to solve: [problem] 1. Write code with inline comments 2. Execute code 3. Verify output correctness 4. If incorrect, debug and re-run 5. Provide final code + results Show all execution attempts.

Copy Generate response, then refine 3 times: V1: Initial response CRITIQUE V1: [identify weaknesses] V2: Improved version CRITIQUE V2: [identify remaining issues] V3: Final polished version Show evolution across versions.

Copy Process these analyses in parallel: THREAD 1: Financial analysis THREAD 2: Market analysis THREAD 3: Competitive analysis THREAD 4: Risk analysis Synthesize findings from all threads. Identify conflicts and synergies.

Copy Analyze documents in multiple languages: - English document: [key points] - Spanish document: [puntos clave] - Japanese document: [重要なポイント] Synthesize in [target language] with cross-references.

Copy Extract data from this chart/image: Output as structured table: | Category | Value | Source Location | |----------|-------|-----------------| Verify extraction accuracy by describing methodology. Flag any ambiguous data points.

Copy Create professional document: FORMAT: [Google Docs / PDF / Markdown] Include: - Proper headers and formatting - Tables where appropriate - Citations in [style] - Table of contents if >5 pages Export as shareable link.

Copy Create comprehensive comparison: Entities: [A, B, C, D, E] Dimensions: [10+ factors] Generate: 1. Comparison matrix 2. Radar chart data 3. Rankings by dimension 4. Overall recommendation Support with data citations.

Copy Analyze trend from historical data: 1. Plot historical timeline 2. Identify inflection points 3. Determine growth rate 4. Project future trajectory (with confidence intervals) 5. Flag anomalies and explain Visualize as chart if appropriate.

Copy For each source used, provide credibility score: [Source Name]: [Score 1-10] - Authoritativeness: [score] - Recency: [score] - Bias assessment: [score] - Verification: [score] Weight conclusions by source credibility.

Copy For each claim, provide confidence: Claim: [statement] Confidence: [0-100%] Reasoning: [why this confidence level] What would increase confidence: [specific info needed] Be honest about uncertainty.

Copy After generating response: FACT-CHECK MODE: For each key claim: 1. Verify with search if possible 2. Look for contradictory evidence 3. Assess source reliability 4. Flag any unverified claims Provide verification report.

Copy Before finalizing response: COMPLETENESS CHECK: ✓ Addressed all parts of question ✓ Provided requested format ✓ Included required details ✓ Met specified constraints ✓ Answered follow-up implications Report: [Items checked] / [Total items]

Copy Review response for bias: 1. Confirmation bias: Am I favoring expected answers? 2. Recency bias: Overweighting recent info? 3. Source bias: Relying too heavily on specific sources? 4. Cultural bias: Assuming specific cultural context? Provide balanced perspective addressing biases found.

Copy Determine optimal response length: Simple query (yes/no): 1-2 sentences Factual query: 1 paragraph Moderate complexity: 2-3 paragraphs Complex analysis: Multi-section response Match length to query complexity, not more.

Copy TEMPLATE: Analysis Report ## Summary [3-sentence overview] ## Key Findings 1. [Finding with evidence] 2. [Finding with evidence] ## Recommendations [Prioritized list] ## Next Steps [Actionable items] Always use this template for analysis requests.

Copy Structure response in layers: LEVEL 1: TL;DR (1 sentence) LEVEL 2: Executive summary (3 bullets) LEVEL 3: Detailed analysis LEVEL 4: Supporting data and appendices User can stop reading at any level.

Copy Citation requirements: - Every factual claim: [Source, Date] - Direct quotes: "Quote" [Source, Page] - Statistics: Data point [Source, Year] - Images/charts: [Original source] No claim without attribution.

Copy Benchmark this response against: - Industry standards - Best practices - Previous solutions - Competitor approaches Provide benchmark score and improvement opportunities.

Copy Generate creative solution with constraints: MUST include: [X, Y, Z] CANNOT use: [A, B, C] Budget: [limit] Time: [limit] Generate 10 ideas that meet ALL constraints.

Copy Solve this problem using analogies from: 1. Nature/biology 2. Architecture 3. Music/art 4. Sports strategy 5. Historical events Extract transferable principles from each domain.

Copy Create 2x2 scenario matrix: Axis 1: [Variable A: High/Low] Axis 2: [Variable B: High/Low] Generate 4 scenarios: 1. High A, High B: [Scenario] 2. High A, Low B: [Scenario] 3. Low A, High B: [Scenario] 4. Low A, Low B: [Scenario] Recommend strategy for each.

Copy Assume perfect success in 3 years. Describe end state in detail. Now work backwards: - What happened in Year 2? - What happened in Year 1? - What happened in Month 6? - What happened in Month 1? Create reverse-engineered roadmap.

Copy Analyze from each stakeholder perspective: CUSTOMER: [needs, concerns, priorities] EMPLOYEE: [needs, concerns, priorities] INVESTOR: [needs, concerns, priorities] REGULATOR: [needs, concerns, priorities] COMMUNITY: [needs, concerns, priorities] Find solution satisfying all stakeholders.

Copy After responding, provide prompt improvement tips: Your prompt was: [effective/unclear] because: - [Specific feedback] To get better results, try: 1. [Specific suggestion] 2. [Specific suggestion] Optimized prompt: [Your rewrite]

Copy Self-assessment rubric: Accuracy: [1-10] - [reasoning] Completeness: [1-10] - [reasoning] Clarity: [1-10] - [reasoning] Usefulness: [1-10] - [reasoning] Overall: [Average] / 10 If any score <7, explain how to improve.

Copy After responding, list knowledge gaps: KNOWN WITH HIGH CONFIDENCE: [X, Y, Z] UNCERTAIN ABOUT: [A, B] DON'T KNOW: [P, Q] To fill gaps, I would need: - [Specific information] - [Specific research]

Copy After solving problem, extract pattern: PATTERN NAME: [Descriptive name] APPLIES TO: [Problem types] DOESN'T APPLY TO: [Exclusions] STEPS: 1. [Generic step] 2. [Generic step] SIMILAR PROBLEMS: [3 examples]

Copy End each response with: REFLECTION: - What worked well in this response - What could be improved - What I learned from this query NEXT TIME: - [Specific improvement to implement] Track improvements across conversation.