Email Intelligence — Structured Data from Unstructured Inputs
Every business has the same inbox problem: invoices that need to be logged, compliance notices that need immediate attention, vendor updates worth tracking, and noise that should disappear. The default solution is a human reading every email and deciding what to do. This automates that decision — and the actions that follow.
Email Router (aCEsXeWi75h6ffV7 — 15 nodes) processes every incoming message. Claude reads the email and classifies it into one of 9 categories. Rules handle the rest — no human triage required for routine volume:
Gmail Trigger ─┐
├─ V4 Rules Classifier → Log to Database → Mark as Read
Webhook ─┘ └─ Has Label? → Apply Gmail Label
└─ Should Archive?
├─ [yes] → Archive Email
└─ [no] → Keep in Inbox
└─ Log Health Check
Error Trigger → Error: Log
Error paths from three nodes merge into one table — any failure in the flow is captured regardless of where it happened.
Invoice Handler (SzwzUJfjjNo6qJyn — 16 nodes) runs daily at 6 AM on whatever the Router flagged as financial. It handles the two formats an invoice arrives in — email body or PDF attachment — and produces the same structured output either way:
Daily 6am / Chat Trigger
└─ Get Financial Emails → Any Emails?
├─ [no] → Done
└─ [yes] → Extract Invoice Data → Has PDF?
├─ [yes] → Download PDF → Upload to Drive ─┐
└─ [no] → Continue ─────────────────────┤
Merge Back
└─ Insert to expenses → Apply Label → Slack Summary
Vendor name, amount, invoice date, due date, IRS category — extracted and written to aos_expenses automatically. Slack summary at the end closes the loop: how many invoices processed, whether anything needs review.
The classifier reads the email rather than matching keywords. That separation — classify once, route by rule — means adding a new email type is one rule addition, not a classifier rewrite. 90%+ of emails process without intervention.
Why it matters: Most businesses are one step away from this — Gmail is already connected, the emails are already arriving. The gap is structured extraction and routing logic. The same architecture applies anywhere unstructured inputs need to become actionable records: support tickets, vendor communications, compliance filings.