|

Overview

I designed and built a modular dispatch automation system around an operational process involving incoming dispatch emails, customer data, project leader information, attachments, appointment details, document storage, SMS communication, and digital signature tracking.

The system connects Outlook, n8n, Microsoft Excel, SharePoint, DocuSign, ClickSend, and Slack into one coordinated automation layer.

Instead of relying on one large workflow, the system is split into multiple workflows for routing, pre-dispatch processing, document handling, dispatch planning, project leader data, SMS communication, DocuSign certificate handling, reminders, and completed document upload.

The challenge

The company handled dispatch operations across three German cities using Outlook, Excel, SharePoint, DocuSign, SMS communication, and team coordination. The process required consistent handling of customer data, project leader details, attachments, folders, appointment records, certificate delivery, reminders, and status updates.

The challenge was to make this process more structured, traceable, and reliable without forcing the operations team to leave the Microsoft 365 tools they already used every day.

What I did

I designed, built, tested, and maintained the automation layer for this dispatch process.

My work included creating the structured Outlook input format, building the n8n workflows, connecting Microsoft 365 services, handling attachments and SharePoint folders, integrating DocuSign and ClickSend, writing JavaScript logic for parsing and validation, and adding Slack notifications for operational visibility.

How it works

The outcome

The system made the dispatch process more structured, traceable, and consistent across multiple regions. It connected incoming Outlook emails, Excel-based operational records, SharePoint document storage, DocuSign certificate workflows, ClickSend SMS notifications, and Slack team updates into one coordinated automation layer.

The biggest impact was automating manual handoffs: requests could be routed automatically, documents stored in the correct customer folders, certificates tracked through their lifecycle, and the team notified when action was needed.

What I learned

Building automation for three operational regions forced me to think in modular workflows, reusable routing patterns, and reliable tracking from the beginning. Subject-line parsing is fragile and needs strict validation. German date formats with cross-month ranges are harder than they look. The biggest lesson: automate the reporting first, because that is what makes the team trust the system.

Proof / Evidence

This case study is based on a multi-workflow automation system. The architecture is documented openly. The full workflow JSON is not published, since it contains customer data, internal SharePoint paths, and credentials.

Public code & documentation

Code excerpt: LLM output sanitizer

An n8n Code node that runs after the Azure OpenAI agent and before any downstream write. Every field returned by the LLM is trimmed, validated, and stripped of zero-width characters. Without this layer, a single hallucinated newline in an email field silently breaks an Excel filter or a DocuSign envelope.

Full version, with date and phone normalization, in snippets/llm-output-sanitizer.js. Eight other snippets in the same repository cover email parsing, batch deduplication, DocuSign envelope ID extraction, SharePoint path sanitization, and scheduled summary aggregation.