I Built an AI Pipeline That Reads Support Emails and Drafts Replies (Here's What Actually Broke, and the Math on Whether It's Worth It)

说明与用途

该流水线自动对支持邮件进行分类和草稿生成，并将通知发送至Slack供人工审核。API成本约为每封邮件$0.01，而人工处理成本约为$1.67。

问题

Small business owners spend hours answering repetitive support emails, many of which are variations of a few common categories.

结果

The pipeline cuts per‑email labor cost from ~$1.67 to ~$0.01 in API spend, while keeping human review. It handles routine high‑confidence cases automatically and routes ambiguous ones to human review, achieving a balanced automation ratio.

工作流

• Pull unread Gmail messages via API
• Extract email body handling multipart/alternative and HTML fallback
• Classify with Claude using structured JSON schema (category, priority, intent, confidence, suggested_reply)
• Route to human based on category and confidence tier
• Generate draft reply in same API call
• Post Slack card with classification, draft, and assignee
• Mark email as read and record result in SQLite
• Human approves or edits draft in Gmail; no auto‑send

证据

The system prompt is about 3,100 characters (~840 tokens), sent on every call along with the email body. At Claude Sonnet's current pricing ($3/MTok input, $15/MTok output), a typical classification-plus-draft call - roughly 1,050 input tokens and 200 output tokens - costs approximately $0.0064 per email.

来源

Industry surveys on support ticket handling time put a straightforward inquiry at somewhere between 3 and 8 minutes of staff time... one manually-answered routine email costs roughly $1.67 in labor.

来源