Skip to content
HoldField

QMS demo · deterministic product illustration

Show how factory signals become controlled QMS recommendations.

A PLC reject spike, with robot and IoT context, becomes a quality signal and an NCR recommendation for QA.

Evidence sealedHuman approvalAudit packet prepared
Open present mode

Physical AI QMS cell

Signal detectedA PLC reject spike, with robot and IoT context, raises a quality signal.
Cameravision captureinspection image + measurement context
PLC robot cellPLC reject spikeA reject-count spike on the PLC, with robot and IoT context around it.
HoldField Podevidence sealedEvidence binding and record preparation
PLC + safetyfinal machine authoritylocal control remains local
robot / motionlocal handlingmotion never waits on QMS routing
Quality engineerhuman decisionQA approval required
QMS recordMachine-to-QMS traceability viewQuality signal · NCR recommendation · Risk review
BoundaryHoldField prepares evidence and routing; Quality engineer approves disposition.
  1. 1Inspection signalcell
  2. 2Machine signalcell
  3. 3HoldField Podpod
  4. 4Quality signalrecord
  5. 5NCR recommendationrecord
  6. 6Audit packetaudit
  • Quality signalPLC reject spike
  • Approval gateQA approval required
  • Readiness view2 complete · 1 open
Signal detected · 1 / 6A PLC reject spike, with robot and IoT context, raises a quality signal.
Use ← / → to step, 1–6 to jump, space to play or pause.

Evidence path

Machine signalpendingReject-count measurementpendingCalibration recordpendingQA signaturepending

QMS records prepared

Records are prepared once evidence is sealed.

QA approval requiredNot yet reached

The system prepared a quality signal and an NCR recommendation from machine evidence.

  • Approver: Quality engineer
  • Human approval required before any disposition.
  • No automatic disposition.

HoldField reads machine signals as data only. Recommendations route to QA; the line stays under local control.

Machine-to-QMS traceability view

Gaps remain
  • Machine-signal evidence — evidence completeSignal, timestamp, and reject count bound to the event.
  • Calibration context — evidence completeCalibration record linked to the station.
  • Root-cause linkage — gap remainsRoot cause is confirmed by a human, not inferred.

A traceability view from machine signal to QMS record — read-only. HoldField never writes to the PLC or commands the line.

Deterministic product illustration. Regulated decisions remain human-controlled. HoldField reads machine signals as data only; QA owns the QMS decision and the line stays under local control.

QMS story: Factory signals become controlled QMS recommendations

Show how factory signals become controlled QMS recommendations.

  1. Step 1 of 6 — Signal detected. A PLC reject spike, with robot and IoT context, raises a quality signal.
  2. Step 2 of 6 — Evidence sealed. The machine signal, timestamp, and calibration context are bound and sealed as evidence.
  3. Step 3 of 6 — QMS record prepared. A quality signal and an NCR recommendation are prepared from the bound machine evidence.
  4. Step 4 of 6 — Human decision required. QA approval is required before the recommendation advances — no automatic disposition.
  5. Step 5 of 6 — Signature recorded. QA records a signature; the recommendation becomes a controlled record.
  6. Step 6 of 6 — Audit packet prepared. An audit trail shows machine-to-QMS traceability; HoldField never commands the line.

Approval: The system prepared a quality signal and an NCR recommendation from machine evidence. Quality engineer must approve. Human approval required before any disposition. No automatic disposition.

Readiness — Machine-to-QMS traceability view: Machine-signal evidence (evidence complete); Calibration context (evidence complete); Root-cause linkage (gap remains). A traceability view from machine signal to QMS record — read-only. HoldField never writes to the PLC or commands the line.

HoldField reads machine signals as data only; QA owns the QMS decision and the line stays under local control.