Say the staffing reality without euphemism, because you run it: your docks are worked by a mix of direct employees, agency labor, and contractor crews, and the mix moves with the season. Peak doubles the agency share. A flu week doubles it again. This is not a failure of your operation — it is how the industry staffs variable volume, and it is not changing.

Now put that fact next to the other one. The dock is where custody actually changes hands. It is where your legal exposure concentrates — claims, detention, OS&D, rework — and where the question that decides those exceptions is always the same: who handled this? Which is precisely the question the industry's event standard cannot record. EPCIS 2.0 defines an event in five dimensions — what, when, where, why, how (§7.2.2) — and no performer is among them; its party fields identify a company, a PGLN, never a person (CBV 2.0 §8.7.1). The check takes ten minutes and the section numbers are in Five Dimensions, No Performer.

The inversion is the whole post: the legally meaningful "who" is asked most exactly where the standard can answer it least.

One shift, reconstructed

Take one cross-dock shift at one of your sites — a mid-size night: 14 inbound trailers, 11 outbound, roughly 60 custody-relevant handoffs (trailer to dock, dock to staging, staging to outbound door). Staff on the floor: 9 direct employees, 6 agency workers on this week's roster, a 4-person contractor crew working two doors under their own supervisor.

Reconstruct the shift as the standard records it, and every one of those 60 handoffs reads identically: the site received; the site staged; the site loaded. Nineteen distinct people touched freight tonight, across three different employment arrangements with three different chains of accountability — and the record contains zero of them. When trailer 7's shipper claims a pallet left complete and arrived crushed, the answer your team can assemble from the record is "our facility handled it," which is exactly the sentence your claims counterparty started with. Everything after it is interviews, camera pulls, and schedule archaeology — billed in your exception hours.

Now the same shift on this record. Every handoff event carries two distinct identities. who is the attested observer — the person who performed the act, resolved through id.org.ai. capturedBy is the warrantor account — who stands behind the capture. The two never collapse, and the three staffing arrangements show why the pair is the minimum viable answer, not a luxury:

  • Direct employee. who: the employee, on their own credential. capturedBy: your site account. The simple case — and even here, two parties.
  • Agency worker. who: the worker, attested. capturedBy: your site account — while the worker's employer is the agency, an org-grain fact the event does not stamp. Which agency held the contract that month is derived at read time from grant chains, so the record stays true after you switch providers.
  • Contractor crew. who: the crew member. capturedBy: the contractor's own account, warranting its crew's scans at your site under the scope you granted. The warrantor question your counsel actually asks — who stands behind a temp's scan? — has a recorded answer with a boundary in it.

Same shift, same 60 handoffs — but now each one names its hands and its warrant, captured at the moment of the act. The full anatomy of the two grains is worked in who Is Not capturedBy; the reason the employer is derived rather than stamped is worked in Reorg-Proof Records. And when the crew at door 9 is a machine instead of a person, the same who field carries the robot — that case is its own post.

Where the money is

The exception costs that land on your P&L — claims you cannot rebut, detention you cannot allocate, rework you cannot trace to a shift — share one root: the unattributed handoff. On your cold-chain lanes the same gap doubles, because custody and condition break at the identical moment: the handoff no record witnessed. Work the math on your own numbers: take last quarter's claims where your defense was "our facility handled it," and ask what fraction settle differently when the answer is an attested observer on each of the three handoffs the freight crossed, with a warrantor behind each one. Then add the quieter line: the exception hours — the interviews and camera pulls — that a query replaces.

Your solutions team feels the same gap from the revenue side: shipper RFPs increasingly carry traceability clauses that ask exactly the who-handled-it question, and a bid that answers with demonstrated attribution rather than policy language is answering a question competitors are still softening. The record that wins those clauses is this one — and because it is a conformant superset of EPCIS 2.0, whose projection validates against GS1's official pinned schema, the answer cites spec sections rather than adjectives. The paperwork side of the same trace — the ASN and PO references your customers' auditors reach for first — is the document-context brief.

What to pilot before you write it into an RFP answer

One dock, one shift pattern, all three staffing arrangements represented — that is the honest pilot. Capture the shift with the two-grain envelope; run your last dispute against the resulting record as a tabletop; count the hours the reconstruction did not take. The measurable is elapsed time from claim to evidence, and it is the number your GM, your counsel, and your next tender all read the same way.

The door here is the get-started interview: your email first, then a short interview that branches on your answers — the 3PL branch asks about your staffing mix, your claims exposure, and the clause in your last RFP you answered more softly than you wanted to. We answer in writing.