Prompts

 

Because Gemini excels at processing multimodal inputs (text, long documents, audio, video, images, and tabular data) and features a massive context window, prompts designed for analysis work best when they leverage these specific capabilities.

Here is a curated list of powerful analysis prompts for Gemini, broken down by domain and use case:

1. Document & Long-Context Analysis

Best for: Whitepapers, academic studies, legal agreements, policy manuals.

• The Executive Brief:

  "Act as an executive researcher. Analyze the attached document and provide: 1) A 3-sentence executive summary, 2) The core methodology or strategy, 3) The top 5 takeaways, and 4) Any hidden risks or unaddressed assumptions."

• Cross-Document Comparative Analysis:

  "Compare Document A and Document B. Map out the points of agreement, key contradictions, and differences in methodology. Present the output in a comparison table followed by an analysis of which perspective is better supported by empirical evidence."

• Adversarial / Red Team Review:

  "Analyze the argument presented in this text. Act as a sharp, critical peer reviewer. Identify logical fallacies, weak evidence, overgeneralized claims, and counterarguments that the author failed to address."

2. Business, Market & Competitive Intelligence

Best for: Strategy teardowns, earning reports, product pitches.

• Earnings Call / Financial Disclosure Deep Dive:

  "Analyze this quarterly earnings transcript. Beyond the headline figures, identify: 1) Shifting priorities or changed guidance, 2) Cautious or evasive phrasing used by leadership during the Q&A, and 3) The most pressing headwind the company faces."

• Competitive Teardown:

  "Perform a competitive analysis between [Company X] and [Company Y] based on their current public offerings. Detail their pricing strategies, target demographics, moat/defensibility, and identify an underserved niche neither is addressing well."

• SWOT + Strategic Initiatives:

  "Conduct a SWOT analysis on [business scenario/product launch]. For each weakness and threat identified, propose one concrete, low-cost operational mitigation."

3. Data & Quantitative Insights

Best for: CSV files, survey responses, raw metric dumps.

• Anomaly & Trend Detection:

  "Analyze this dataset. Identify: 1) The top 3 macro trends over time, 2) Any significant anomalies or outlier data points, and 3) Potential causal relationships or correlations worth investigating further. Provide hypotheses for why these outliers occurred."

• Customer Sentiment & Categorization:

  "Here is a batch of raw customer feedback/reviews. Categorize the feedback into thematic buckets (e.g., UI/UX, Pricing, Reliability, Feature Requests). Quantify the prevalence of each theme and highlight the top 3 recurring friction points causing churn."

• Cohort & Segmentation:

  "Review the attached user behavior metrics. Propose three distinct behavioral customer segments based on usage frequency, engagement depth, and feature adoption. What retention strategy would you apply to each group?"

4. Multimodal Analysis (Images, Charts, Video, Audio)

Best for: Infographics, UI mockups, lecture recordings, diagrams.

• UI/UX & Visual Teardown (Upload image/mockup):

  "Analyze this user interface screenshot. Evaluate it based on visual hierarchy, accessibility (contrast, typography), and cognitive load. List the top 3 friction points a first-time user might face and propose concrete layout fixes."

• Chart & Infographic Fact-Check (Upload chart/graph):

  "Analyze this data visualization. 1) Summarize what the graph shows, 2) Check for misleading visual tactics (truncated axes, disproportionate scales, selective date ranges), and 3) State the most statistically honest takeaway from the data."

• Long-Form Video/Audio Breakdown (Upload media or link YouTube):

  "Analyze this recorded panel/meeting. Create a chronological breakdown with timestamps for every major topic shift. For each section, summarize the key debate points, who argued what, and what action items were agreed upon at the conclusion."

5. Technical, Code & Architecture Analysis

Best for: Codebases, API specs, architecture diagrams.

• Codebase Audit & Vulnerability Scan:

  "Review the attached code. Analyze it for: 1) Security vulnerabilities (e.g., injections, race conditions, unsanitized input), 2) Time/space complexity bottlenecks (O(n) analysis), and 3) Maintainability/code smell. Suggest refactored snippets for the two most critical issues."

• Architecture Trade-Off Evaluation:

  "Evaluate the attached system architecture. What are the primary single points of failure (SPOF)? How will this system behave under a 10x traffic spike? Propose two structural improvements to enhance fault tolerance without significantly increasing cloud infrastructure costs."

6. Creative & Narrative Textual Analysis

Best for: Scripts, marketing copy, speeches.

• Tone & Rhetoric Deconstruction:

  "Analyze the attached speech/op-ed. Examine the rhetorical devices used (ethos, pathos, logos, framing, repetition). Who is the intended primary and secondary audience, and what emotional levers is the speaker attempting to pull?"

• Narrative Pacing & Character Arc:

  "Read this draft chapter/script. Analyze the scene pacing, subtext in the dialogue, and character agency. Does the protagonist drive the plot here, or are they purely reactive? Provide specific line-level notes."

Pro-Tip: The Anatomy of a High-Performing Gemini Analysis Prompt

To get the sharpest output, structure your prompts using this four-part formula:

1. Role/Persona: "Act as a senior data analyst / compliance attorney / chief strategist..."
2. Context + Artifact: Attach the file or paste the text, explicitly noting what the document represents.
3. Analytical Framework: Specify how to analyze it (e.g., first-principles thinking, comparative evaluation, SWOT, root cause analysis).
4. Constraint & Format: "Limit your response to 500 words, prioritize actionable insights over summaries, and output findings in a Markdown table."