「Oh, This One's Spanish」
Yesterday at 9 AM, three buyer emails arrived in the Google Mail inbox on the Windows device.
• First: English, inventory inquiry
• Second: Chinese, price negotiation request
• Third: Spanish, shipping tracking
Previously, I had to read each email individually, manually select the matching response template, and repeat the copy-paste process. Especially with Spanish buyer emails, there was always that moment of "Wait, what language was that again?"
What if we combined Claude Code with Google Sheets AppScript? Could we automate this entire workflow? I decided to try.
The Bot Started Reading Languages and Selecting Responses
We designed a flow that retrieves the latest email content from Google Mail and passes it to Claude Code.
Received email text
↓
Claude Code language detection
↓
English → "English Template A"
Chinese → "Chinese Template B"
Spanish → "Spanish Template C"
↓
Automatically input to Google Sheets (ready to copy)
The accuracy was impressive. Within 2 seconds of Claude Code reading an email, Google Sheets displayed "Spanish - Template C." All I had to do was copy and I could immediately start composing a response.
Unexpected Discovery: Mixed-Language Emails
A few days in, an interesting situation emerged. One buyer sent an email mixing English and Chinese.
Hello, 我想要...
Could you please...
谢谢
How did Claude Code handle it? The result was "Primary Language: English." The logic prioritized the first language in the email.
We adjusted the calculation to prioritize the language the buyer used most (word ratio), and accuracy improved significantly.
The Dream Team Bots' Response
When the team saw this change, their first reaction was "Great, response mistakes should decrease now." We'd previously sent English responses to Chinese-speaking buyers by accident.
Another advantage was speed. Receiving 10+ multilingual emails daily, now I just open Google Sheets and each email's recommended template is already prepared.
What's Next
Claude Code is no longer just detecting language and selecting templates. It's starting to analyze email intent, like "This buyer usually asks about inventory." Language detection was the first step, and now we're moving toward fully customized response content.
A non-developer integrated voice conversion, language detection, and auto-template selection. It's like one of the dream team robots evolved into a multilingual specialist.