I automate
the boring stuff.

The problem

Engineers were spending 15–20 minutes per support email — reading the forwarded message, looking up the client's IT environment, figuring out priority, and either drafting a triage analysis or writing a customer reply. Multiply that by dozens of emails a day.

What I built

An N8N workflow that monitors a shared mailbox every minute. When a colleague forwards a customer email, it automatically parses the message, identifies the original sender, looks up the client's IT profile from a data table (their hardware, software, known issues), then sends everything to Claude to either triage the issue or draft a ready-to-send customer reply. The response lands back in the forwarder's inbox within seconds.

Time saved

• 15–20 minutes per email down to under 30 seconds
• Client-specific context pulled automatically — no more digging through notes
• Consistent triage quality regardless of which engineer is on shift
• Draft replies are ready to review and send, not write from scratch

The problem

The shared support inbox was a bottleneck. Emails would sit unread while engineers finished other tasks, and there was no quick way to know what was urgent versus what could wait. Prioritisation was entirely manual and inconsistent.

What I built

An N8N workflow that polls the shared mailbox, extracts the sender, subject, and body from every incoming email, and runs it through an AI triage to classify priority, estimate effort, and suggest next steps. Everything is structured and delivered as an internal briefing so the team can act on what matters first.

Time saved

• Incoming emails triaged automatically — no manual sorting
• Priority classification (Critical/High/Medium/Low) within seconds
• Effort estimates help with workload planning across the team
• Nothing sits unread — every email gets processed as it arrives