← Back to recipes
Save

Messy Excel Sales Data to Executive Visual Report

Every Monday I'm staring at ten thousand order rows needing an exec dashboard by 10am — my pivot table alone takes until 9:30.

The person who submitted this, in their own words

Transforms thousands of raw transaction rows into an interactive executive dashboard with KPI anomalies.

When it won't work

  • Anything needing finance-grade reconciliation — check the numbers with finance first
  • Under a few dozen rows: a plain chart is faster than this whole flow

ExcelSpreadsheetChartReporting

  • TypeTables & data
  • ForOperations
  • Time3 min
  • EffortEasy
  • CostFree quota
  • CheckedCopyAI tested

Before you start

  • A spreadsheet of raw sales or orders (CSV/Excel)
  • Access to any modern LLM chat interface

Step by step

Input Raw Data Samples

Export your Excel sheet as CSV or copy the first 30 rows of column headers and sample data into your clipboard.

Watch out · For large files over 50MB, copy just the table schema and let AI write a local python script.

Run Parameterized Business Analyst Prompt

Adjust the target sector and audience below, then copy the prompt into your AI conversation:

You are a Fortune 500 Senior Commercial Analytics Director.
Analyze the uploaded transaction data for the [{{industry}}] sector prepared for [{{audience}}].

Please output:
1. Executive KPI Summary: Total Revenue, Net Margin, AOV, MoM fluctuations in a clean Markdown table.
2. Anomaly Attribution: Identify top 3 underperforming categories and root causes.
3. Growth Opportunities: High-margin breakout items and customer segment characteristics.
4. Actionable Next Steps: 3 zero-budget initiatives for the next sprint.

Generate HTML / PDF Dashboard

Ask the model: "Convert the above analysis into a modern standalone HTML dashboard with responsive metric cards and bar charts." Open the file directly in Chrome.

How to know you did it right

  • Calculated totals match original raw spreadsheet sums
  • Specific anomalies identified with actionable recommendations

Who checked it

CopyAI tested · Tested on 2026-11-01 · 1428 made it

Original source https://www.anthropic.com/news/contextual-retrieval

Where it usually breaks

把整张表粘进去,前端一卡就失败。超过 50MB 只贴表头和前 30 行。

Starter files

Did you make it?

Tell us how it went. If it broke, say where — that is how we fix the recipe.

How did it go