The problem
Shared inboxes mix time-sensitive issues, invoices, and routine mail. A naive AI agent can route quickly but may also make a confident-looking mistake with no review trail.
Case study 01 · n8n · AI workflow
AI email routing with a human decision boundary.
The workflow classifies an incoming email as urgent, invoice, or routine—but AI never gets the last word when confidence is low.
Public demonstration. Accounts and credentials are not included; the workflow uses buyer-supplied services and synthetic examples.
0.70human-review threshold
3explicit routing categories
1audit path for every branch
// try the browser demo
Choose a synthetic message
This browser demo uses fixed examples to show the decision boundary. It does not call an AI service or connect to an inbox.
The low-confidence example stops at human review instead of guessing.
—category
—confidence
—route
Ready. Select a message and run the triage walkthrough.
Public demonstration · synthetic messages · no credentials · no paid-client data
// system flow
Email trigger→Structured AI classification→Confidence gate→Explicit branch→Audit log
The control design
The model must return a fixed JSON schema: category, confidence, and one-sentence reason. Confidence below 0.70 bypasses automation and alerts a human-review queue.
Deterministic actions
Urgent messages alert the team, invoices move through filing and ledger steps, and routine messages are labeled. The workflow—not the model—chooses the allowed action.
Handover boundary
The importable n8n JSON contains no credentials. Operators connect their own inbox, model, team-chat, storage, and sheet accounts after reviewing the nodes.
// proof available
email → {category, confidence, reason} → confidence gate → route → decision logThe complete workflow is available as an importable starter kit. The same pattern can be adapted to lead intake, support queues, document review, or approval routing without claiming autonomous judgment.
Need this mapped to your inbox?
Send the categories, exception examples, tools, and desired human-review rule.
[email protected]