AI Email Orchestrator

Two types of Email Orchestrator built

The problem

One shared inbox. Job applications, client family inquiries, caregiver call-offs, scheduling changes, invoices — all landing in the same place, all waiting for a human to read them and forward them to the right person.

The dangerous ones were the time-sensitive ones: a caregiver calling off a shift at 6 AM shouldn't wait until someone checks email at 9.

What I built

An email orchestration workflow in n8n that acts as a full-time triage assistant:

• Intent classification — every inbound email runs through a keyword engine that identifies what it is: application, client inquiry, call-off, scheduling change, billing, or spam.
• Smart routing — each intent has a destination. Call-offs alert the on-call coordinator instantly. Applications flow into the recruitment pipeline. Billing goes to the bookkeeper.
• Automated first responses — applicants get an acknowledgment with next steps immediately, so good candidates don't go cold.
• Escalation rules — anything the engine can't classify confidently goes to a human review queue rather than being guessed at.

Why it matters for your agency

Your inbox is a queue of business events wearing an email costume. Treating it as one means urgent things get urgent handling and routine things handle themselves.

In the demo below, pick a sample email (or write your own) and watch it flow through the actual routing logic.

Under the hood

The production workflow runs in n8n with a Gmail trigger. Classification uses a keyword ruleset engine — deterministic and auditable, no black-box AI deciding where a client complaint goes. Routing actions include Slack alerts, SMS via the agency's CRM, label + folder moves, and template replies. The demo on this page replays the same decision tree with simulated data.