VARO. Automated n8n Invoice Extractor.Colombia.

I need a well-structured n8n workflow that watches my Outlook inbox for new supplier invoices, opens each message and its attachments (PDF, XML DIAN, .eml and even ZIP bundles that might contain them), and extracts the key accounting fields—supplier, NIT, date, item lines, totals, CUFE, and any other standard tags. The data must be written in real time to both an Excel file (local or OneDrive is fine) and a Google Sheet, keeping identical column order so my accountant can pivot between tools without re-formatting.

Core expectations
• Use standard n8n nodes (Outlook, IMAP, Merge, Function, PDF reader, XML parser) or lightweight custom JS when a built-in node falls short.
• Map the DIAN XML schema reliably; if the file is a PDF, apply text or ZUGFeRD-style extraction—whatever gives the fields with the highest accuracy.
• Handle ZIP archives by iterating through their contents before parsing.
• Flag unreadable or incomplete invoices, send a brief error note to a Slack/Teams webhook, and continue processing the rest.
• Comment every node so I can tweak credentials or field names later without reverse-engineering your logic.

Deliverables
1. Exported .json of the finished n8n workflow.
2. A sample Excel workbook and Google Sheet populated from at least three test invoices (one PDF, one XML DIAN, one inside a ZIP).
3. A short markdown or PDF guide that covers installation, required credentials, and how to add new suppliers or columns.

Acceptance criteria
• When I drop new invoices into Outlook, both spreadsheets update within one minute and show 100 % correct values for supplier, NIT, date, items, totals and CUFE.
• Workflow runs without manual intervention for 24 hours, logging any errors but never halting on a single failure.

If you’ve built similar automations in n8n, especially around Latin American e-invoicing or DIAN, let me know—your approach will likely meet my needs fastest.