The Problem: Format Chaos
Every week, overseas buyers send order data. But the same information arrives in wildly different formats.
Some buyers use A4 landscape orientation, others portrait / headers are in different places / date formats vary / quantity units differ.
Before: Manual Conversion
• Manually open and check every received file
• Copy and paste data into our system format by hand
• Reformat dates and numbers
• Time per order: about 8 minutes
• Error rate: High (easy to miss or confuse)
The Turning Point: Claude Code Auto-Conversion
We set Claude Code to read raw files arriving in our Google Sheet. Everything else followed naturally.
# What Claude Code does
1. Auto-detect buyer file structure
2. Identify header locations
3. Recognize data types (numbers, dates, text)
4. Convert to our standard format
5. Automatically input converted data into main sheet
After: Automated Conversion Workflow
Actual process:
1. Buyer file arrives in Windows device inbox
2. Automation script detects file
3. Claude Code executes instantly
4. Standard format conversion completes in 3 seconds
5. Data auto-inputs to Google Sheet
The Unexpected Discovery
Here's what surprised us most.
Claude Code started finding problems while converting.
• Detects duplicate orders accidentally entered by buyers
• Auto-flags rows with unusual number formats
• Auto-alerts to currency unit mismatches
Data errors we could never spot with our eyes now appear on the checklist.
Numbers Tell the Story
| Metric | Before | After |
|--------|--------|-------|
| Time per order | 8 minutes | 3 seconds |
| Monthly volume (100 orders) | 800 minutes (13 hours) | 5 minutes |
| Error rate | ~12% | 0.3% |
| When errors are found | Before shipping | At entry |
What We Learned
The biggest gain wasn't speed. We gained reliability.
The more diverse buyer formats become, the higher the risk of hidden errors. Claude Code isn't just unifying formats. It became our data quality control inspector.
Now we don't blindly trust incoming data. We let Claude Code verify first.