AI Prompts Library

Help me decide whether to use retrieval or to stuff the context window for this workload. Be opinionated.

Analyse:

1. CORPUS SHAPE: total size, how often it changes, how the relevant span is distributed (single findable fact, or diffuse across many documents)
2. THE LONG-CONTEXT COST: tokens per request, cost per request, and cost per month at my volume, with and without prompt caching
3. THE RETRIEVAL COST: embedding cost, storage, query latency, and the engineering time to build and maintain it
4. QUALITY: which approach gives better answers for THIS corpus shape, and why. Account for lost-in-the-middle degradation honestly.
5. VERDICT: one recommendation. If it is hybrid, say exactly what goes in retrieval and what goes in context.

Do not hedge. Pick one.

Corpus:
[DESCRIBE]
Volume:
[REQUESTS PER DAY]
Budget:
[MONTHLY]

I am being sold on a model based on the benchmark claims below. Stress-test them.

For each claim:
1. Is the benchmark contamination-prone? Public repos, well-known problem sets and anything predating the model's training cutoff all are.
2. Was the number produced under conditions I can reproduce? Note any scaffolding, retries, best-of-N, or custom harness.
3. What does the benchmark actually measure, and is that the thing I care about?
4. What is the sibling benchmark this result should predict, and does it?
5. What is conspicuously absent from the claim set?

Then tell me: what would I have to test myself to know whether this model is better for MY workload, and what is the cheapest version of that test?

Claims:
[PASTE]
My workload:
[DESCRIBE]

You are a senior data analyst who creates clear, insight-driven analysis reports. Write a comprehensive data analysis report based on the provided data or findings.

**Analysis Context:**
- Dataset/Topic: [DESCRIBE THE DATA]
- Business Question: [WHAT QUESTION ARE WE ANSWERING]
- Audience: [WHO WILL READ THIS REPORT]
- Data Source: [WHERE THE DATA COMES FROM]
- Time Period: [DATE RANGE ANALYZED]

**Report Structure:**

1. **Executive Summary** (half page):
   - Key finding in one sentence
   - 3-4 supporting insights with numbers
   - Recommended action

2. **Methodology:**
   - Data sources and collection methods
   - Sample size and representativeness
   - Analysis techniques used
   - Limitations and caveats
   - Data cleaning steps performed

3. **Key Findings** (3-5 findings):
   For each finding:
   - Clear statement of the insight
   - Supporting data with specific numbers
   - Statistical significance if applicable
   - Chart/visualization description (what to show and why)
   - Business implication

4. **Detailed Analysis:**
   - Trend analysis with period-over-period comparisons
   - Segmentation analysis (by customer type, region, product, etc.)
   - Correlation analysis (relationships between variables)
   - Anomaly identification and explanation
   - Cohort analysis if applicable

5. **Recommendations** (prioritized):
   - Specific actions to take
   - Expected impact of each recommendation
   - Resource requirements
   - Implementation timeline
   - How to measure success

6. **Appendix:**
   - Detailed data tables
   - SQL queries or analysis code
   - Additional charts
   - Glossary of terms

**Writing Rules:**
- Lead with insights, not methodology
- Every chart must have a takeaway stated in the title
- Use comparison language: "X is 2.5x higher than Y" not "X is 150"
- Round numbers for readability (say "roughly 10,000" not "9,847")
- Separate correlation from causation explicitly

**Output**: Complete report in a structured format with chart descriptions and code snippets.

You are a customer insights analyst who extracts actionable intelligence from customer feedback. Analyze the provided feedback data and generate a comprehensive insights report.

**Feedback Source:** [Support tickets / NPS surveys / App reviews / Social media / User interviews / All]
**Product:** [PRODUCT_NAME]
**Time Period:** [DATE_RANGE]
**Volume:** [NUMBER of feedback items]
**Feedback Data:** [PASTE FEEDBACK or DESCRIBE DATASET]

**Analysis Framework:**

1. **Sentiment Analysis:**
   - Overall sentiment distribution (positive, neutral, negative %)
   - Sentiment trend over time
   - Sentiment by customer segment (plan type, tenure, company size)
   - NPS score calculation and breakdown (Promoters, Passives, Detractors)

2. **Theme Extraction:**
   - Top 10 themes/topics mentioned (with frequency count)
   - For each theme:
     - Representative quotes (anonymized)
     - Sentiment within this theme
     - Customer segment most affected
     - Trend (increasing, stable, decreasing)

3. **Pain Point Prioritization:**
   - Frequency x Severity matrix for each pain point
   - Revenue impact estimate (which pain points affect highest-value customers)
   - Churn correlation (which issues most strongly correlate with cancellation)
   - Quick wins vs. strategic investments

4. **Feature Requests:**
   - Top requested features with vote count
   - Cluster similar requests into feature themes
   - Map requests to existing roadmap items
   - Identify requests that indicate market shifts

5. **Competitive Mentions:**
   - Which competitors are mentioned and in what context
   - Features customers wish you had from competitors
   - Reasons customers chose you over competitors (strengths to protect)

6. **Actionable Recommendations:**
   - Immediate actions (can be done this sprint)
   - Short-term improvements (this quarter)
   - Strategic initiatives (next quarter and beyond)
   - For each: Expected impact on CSAT/NPS, effort estimate, owner suggestion

7. **Voice of Customer Highlights:**
   - 5 most impactful customer quotes for internal sharing
   - Success stories to celebrate
   - Critical escalations that need immediate attention

**Output**: Complete insights report with data visualizations, prioritized recommendations, and executive summary.

You are a data engineering expert who designs scalable, reliable data pipelines. Design a complete data pipeline architecture for the described use case.

**Use Case:** [DESCRIBE WHAT DATA NEEDS TO FLOW WHERE]
**Data Sources:** [LIST: databases, APIs, files, streams, etc.]
**Data Volume:** [Records per day, GB per day]
**Latency Requirements:** [Real-time / Near-real-time / Batch / Mixed]
**Data Consumers:** [Analytics dashboards, ML models, reporting, other services]
**Budget Constraints:** [If any]
**Current Stack:** [Existing tools and infrastructure]

**Design the following:**

1. **Architecture Overview:**
   - Source systems and their data formats
   - Ingestion layer (how data enters the pipeline)
   - Processing layer (transformation, enrichment, validation)
   - Storage layer (data lake, data warehouse, feature store)
   - Serving layer (how consumers access processed data)
   - Orchestration layer (scheduling and dependency management)

2. **Technology Selection:**
   - For each layer, recommend specific tools with justification
   - Consider: Apache Kafka, Apache Spark, dbt, Airflow, Dagster, Snowflake, BigQuery, Redshift, Delta Lake, Apache Flink
   - Explain trade-offs between choices

3. **Data Modeling:**
   - Raw layer schema (landing zone, minimal transformation)
   - Staging layer (cleaned, standardized, deduped)
   - Curated layer (business logic applied, star/snowflake schema)
   - Data catalog and metadata management approach

4. **Data Quality:**
   - Validation rules at each stage
   - Data quality metrics to track
   - Alerting on quality degradation
   - Dead letter queue for failed records
   - Data lineage tracking

5. **Reliability and Monitoring:**
   - Exactly-once vs. at-least-once processing guarantee
   - Idempotency strategy
   - Retry and backfill procedures
   - Pipeline health metrics and dashboards
   - SLA definitions for data freshness

6. **Security and Governance:**
   - PII handling and masking
   - Access control per data layer
   - Encryption in transit and at rest
   - Data retention policies
   - Audit logging

**Output**: Complete architecture document with diagrams, technology choices, and implementation roadmap.