SYNTHETIC - NO CUSTOMER ACCESS

BUILD-AUTOMATE

Lead-routing workflow simulation for a synthetic service business

A static process replay showing how each synthetic request is classified, routed, escalated, and handed to a human without connecting to any real system.

The job

Customer problem

A small field-service company loses time when new requests are copied between an inbox, a spreadsheet, and an operator chat without consistent urgency or ownership rules.

Synthetic inputs

Synthetic inquiry queue

Six fictional requests covering routine service, an urgent outage, an unsupported region, a duplicate, a missing phone number, and an after-hours request.

Routing policy

Urgent outages require human escalation; unsupported regions are declined; duplicates merge; incomplete requests enter a clarification queue.

Operator roster

Three fictional operators with service regions, working hours, and maximum active assignments.

What the AI did

  1. Normalized every inquiry into the same typed request shape.
  2. Applied deterministic urgency, region, completeness, duplicate, and capacity rules.
  3. Produced a proposed owner and reason for every route.
  4. Separated human approvals and exceptions from routine assignments.
  5. Rendered the state transitions and minimum fields a later integration would require.

Findings and deliverables

Acceptance checks

PASS

Every inquiry receives one visible terminal or waiting state

All six synthetic requests end as proposed assignment, clarification, merged duplicate, decline, or human escalation.

PASS

Urgent and after-hours work cannot auto-dispatch

Both cases stop at a human approval state with the reason displayed.

PASS

The proof performs no external action

No email, chat, calendar, CRM, or dispatch system is available to the synthetic run.

What remains unproven

Next access request

None for the first workflow review. After approval, request only an approved field list and read-only sample schema before discussing any connector.