Platform

Four surfaces. One record.

Relay is built around the work that actually crosses a site boundary: intake, the clock, the pass, and the view from HQ.

01 / Intake

Orders land once, then route.

Inbound work becomes a ticket with a site, type, and promised window. Dispatchers assign an owner; operators see only the queue that belongs to their floor.

  • CSV, webhook, or manual create
  • Site + order-type routing rules
  • Idempotent ticket keys so reprints don’t duplicate
WhenThenOwner
type = kitPlant A · Dock 2Incoming
type = replenishPlant B · PickL. Park
hazmat = trueHold + SDSCompliance
sla < 30mPage dispatcherOn-call

02 / Clocks

SLAs that change color before they fail.

Each order type has a window. Relay computes remaining time against the site’s local clock, then moves the ticket from on-track to at-risk to breached.

  • Per-type windows, pauses on customer hold
  • At-risk threshold you set (default 20%)
  • Escalation to the next role, not a group inbox
TypeWindowLeftState
Same-day pack45m14mOn track
Exception20m3mAt risk
Carrier cut30m0Breached

03 / Pass

Handoffs you can audit.

A handoff is a first-class object: from-shift, to-shift, blockers, and the tickets that still own the floor. Close it only when the incoming lead acknowledges.

  • Required fields per site
  • Open tickets attach automatically
  • Acknowledgement timestamp on the record
FieldValue
FromNights · A. Ruiz
ToDays · M. Chen
OpenORD-2398 at risk, HAN-882 QC
BlockerLabel printer 4 down — parts AM
Ackpending

04 / Sites

The same table, a different floor.

HQ sees every site. A supervisor sees one. Switch plants and the queue, owners, and clocks update — not a new login, not a copied sheet.

  • Plant-level filters on every list
  • Shared ticket IDs across the network
  • Read-only HQ role for the weekly review

Plant A · Austin Assembly

IDWorkSLAOwnerState