校验销售报表 · Report Validator

它解决什么问题

销售报表的数字错一点,决策就偏一大截:合计对不上、环比口径不一致、分类漏项、异常值没人解释。直接让 AI 检查,常只挑拼写错误,抓不住数字逻辑。

这个技能做什么

把报表按「合计校验 → 口径校验 → 分类占比 → 异常值」四层检查:核对加总与明细是否一致、环比口径是否统一、占比是否合理、异常波动是否有解释;输出可信度评级与待人工确认清单。

为什么比裸提示词强

• 查逻辑:不止看格式,还看数字勾稽。
• 可追溯:每个问题指向具体单元格。
• 带评级:报表能不能直接进会议一目了然。

典型用例

• 周会/月会前校验销售报表。
• 审计历史报表的口径一致性。
• 新人做的报表做质量把关。

局限

需要你提供报表数据;无法自动核对原始系统数据,标注「待确认」处需人工查证。

English · 英文介绍

The problem

If the numbers in the sales report are a little wrong, the decision-making will be a lot wrong: the totals are not correct, the month-on-month caliber is inconsistent, the categories are missing, and there is no explanation for the outliers. Let AI check it directly, and it often only picks out spelling errors and fails to grasp numerical logic.

What it does

Check the report at four levels: "Total Verification → Calibration Verification → Category Proportion → Abnormal Value": Check whether the total and details are consistent, whether the month-on-month caliber is unified, whether the proportion is reasonable, and whether there are explanations for abnormal fluctuations; output credibility ratings and a list to be manually confirmed.

Why is it better than naked prompt words?

• Check the logic: not only look at the format, but also look at the number.
• Traceability: Each question points to a specific cell.
• With rating: It is clear at a glance whether the report can be directly entered into the meeting.

Typical use cases

• Verify sales reports before weekly/monthly meetings.
• Audit the caliber consistency of historical statements.
• Check the quality of reports made by new employees.

Limitations

You are required to provide report data; the original system data cannot be automatically checked, and manual verification is required where "to be confirmed" is marked.