SuperAPI AI Directory

insight-generation

Data & analytics

View repository

Updated 10-16 08:46

About this skill

Load context

On first use: Read ../../CLAUDE.md to obtain scenario-level configuration (business context/historical benchmarks/industry benchmarking data).


/insight-generation — Insight Generation

Examples

→ Example: The user says "This month's operational data is out, help me generate a few core insights that management should pay attention to." The system should invoke this skill to distill key insights from multidimensional data.

→ Example: The user says "Help me add a few strategic-level insights to the quarterly report, not just financial figures." The system should invoke this skill to generate strategic-level business insights.

→ Example: The user says "For the new market entry strategy, help me provide a few key decision-support insights from a financial perspective." The system should invoke this skill to generate decision-support insights.

Step 1: Data Acquisition and Quality Assessment

Data extraction:

□ 数据来源:[BI 系统/ERP/财务系统/CRM]
□ 查询条件:期间 [YYYY-MM 至 YYYY-MM],维度 [客户/产品/渠道/区域]
□ 记录数:[X] 条
□ 数据字段:[列表]

Data quality check:

□ 完整度:[X]% — 缺失 [字段/记录] 的原因
□ 一致性(跨系统):[✅ 一致 / ⚠️ 存在差异]
□ 时效性:数据截止 [YYYY-MM-DD] — [✅ 及时 / ⚠️ 延迟 X 天]
□ 总体质量评估:[✅ 高 / ⚠️ 中 / 🔴 低]

Step 2: Descriptive Analysis

Key metrics overview:

| 指标 | 本期值 | 同比 | 环比 | 达成率 |
|------|--------|------|------|--------|
| [指标1] | [X] | [±X]% | [±X]% | [X]% |
| [指标2] | [X] | [±X]% | [±X]% | [X]% |

Distribution analysis:

□ 集中度:
  → Top 3 贡献占比:[X]%(客户/产品/渠道)
  → 是否过度集中:[✅ 否 / ⚠️ 是,集中度 [X]%]

□ 分布形态:
  → [指标] 分布:[正态/偏态/双峰]
  → 区间分布:[X]% 的值落在 [区间]

Step 3: Pattern Recognition

Trend patterns:

□ 增长形态:
  → 线性增长:月均增 [X]
  → 季节性波动:Q4 峰值 [X],Q1 低谷 [X],波幅 [X]%
  → 阶梯式增长:[日期] 前后台阶 [X]
  → 平台期:[日期] 起增长放缓

□ 主要驱动因素(归因):
  → 增长 [X]% 中,[因素A] 贡献 [X]%,[因素B] 贡献 [X]%

Anomaly patterns:

□ 异常点检测:
  → [日期/客户/产品] 的 [指标] 值为 [X],偏离趋势 [±X]%(阈值 [X]%)
  → 偏离方向:[高于/低于] 预期
  → 异常类型:[一次性/系统性]

□ 异常根因初步判断:[描述]

Correlation analysis (if applicable):

□ 指标间相关性:
  → [指标A] 与 [指标B] 相关系数 [X](强正/强负/弱)
  → [指标C] 滞后 [指标D] [X] 期

Step 4: Insight Distillation

Structured insight output:

| # | 洞察描述 | 类型 | 支撑数据 | 置信度 | 业务影响 | 可行动性 |
|---|---------|------|----------|--------|----------|----------|
| 1 | [描述] | [模式/异常/机会] | [X] | [高/中/低] | [高/中/低] | [高/中/低] |
| 2 | [描述] | [模式/异常/机会] | [X] | [高/中/低] | [高/中/低] | [高/中/低] |

洞察类型定义:
  模式(Pattern):反复出现的规律性现象
  异常(Anomaly):偏离正常范围的异常点
  机会(Opportunity):未被充分利用的增长点

Detailed explanation of Top 3 insights:

洞察 1:[标题]
□ 描述:[具体说明]
□ 发现依据:[数据/现象]
□ 成立条件:[在什么前提下成立]
□ 潜在风险:[洞察可能不成立的情况]
□ 置信度:[高/中/低] — 依据 [样本量/一致性/逻辑性]

洞察 2:[标题]
...

洞察 3:[标题]
...

Step 5: Recommendation Transformation

Insight → Recommendation mapping:

| 洞察 | 转化为的行动建议 | 预期效果 | 实施难度 | 优先级 |
|------|----------------|----------|----------|--------|
| [洞察1] | [建议] | [量化效果] | [高/中/低] | [高/中/低] |
| [洞察2] | [建议] | [量化效果] | [高/中/低] | [高/中/低] |

Recommendation feasibility assessment:

□ 建议 1:[描述]
  → 所需资源:[人力/资金/时间]
  → 实施周期:[X] 周
  → 收益成本比:[X]
  → 主要风险:[描述]
  → 综合可行性:[✅ 可行 / ⚠️ 需进一步评估]

Step 6: Generate Insight Report

═══════════════════════════════════════
洞察生成报告
分析主题:[主题]
数据期间:[YYYY-MM 至 YYYY-MM]
生成时间:[YYYY-MM-DD HH:MM]
═══════════════════════════════════════

【分析概览】
□ 数据质量:[✅ 高 / ⚠️ 中 / 🔴 低]
□ 关键指标:[X] 个
□ 主要发现:[X] 条

【核心洞察】(按影响力排序)

🔵 洞察 1:[标题]
   类型:[模式/异常/机会] | 置信度:[高/中/低] | 影响:[高/中/低]
   描述:[详细说明]
   建议行动:[描述]

🔵 洞察 2:[标题]
   ...

🔵 洞察 3:[标题]
   ...

【次要洞察】
□ [洞察4]:[描述] — 置信度 [中/低]

【建议行动】(按优先级)
1. [优先级最高,建议立即执行]
2. [优先级次之,建议本周执行]
3. [长期改进项]

【待验证假设】
□ [假设1] — 验证方法:[描述]
□ [假设2] — 验证方法:[描述]

═══════════════════════════════════════
置信度:[✅ 高 / ⚠️ 中 / 🔴 低]
洞察总数:[X] 条 | 高影响:[X] 条 | 可行动:[X] 条
═══════════════════════════════════════

Escalation Trigger Conditions

  • Overall insight confidence is low (mostly medium/low confidence)
  • Systematic anomalies found (involving more than [X] data points)
  • Insights involve correlated impacts across multiple business domains
  • Recommended actions involve cross-department coordination or major resource investment
  • Low data quality (assessed as 🔴 Low) limits insight reliability

Finance Skills — insight-generation atomic skill