SPREAD's deployment at the customer's Central-European assembly plant for the new electric-architecture ramp is Production Inspector. It ingests wiring-harness data, BoMs (bills of materials), communication diagrams, and 3D models, organized as a VIN-specific (vehicle identification number) Product Twin. The rework operator at the workstation scans the VIN, the system filters the entire dataset to that vehicle's actual configuration, and the diagnostic question the operator faces is answered against the vehicle in front of them, not a notional one. Live since 2025 across the customer's Central-European plant, with adjacent-plant onboarding in May 2026 and an additional plant queued on the same shared architecture.
The change that doesn't show up on a financial statement is who is qualified to answer technical questions at the line.
The new electric architecture ramp is a complexity ramp
The customer's Central-European plant is the assembly line for the new electric architecture: a centralized compute platform with substantially higher software content per vehicle than the platforms it replaces. Diagnostic errors during ramp combine ECU (electronic control unit) faults, signal anomalies, and dependency effects in ways that don't decompose cleanly. A signal-bus anomaly might point to a hardware defect, a configuration mismatch, a software regression in a recently-shipped variant build, or a wiring-harness assembly error specific to that vehicle's customer-spec configuration.
Historically, this distinction was made by the senior engineer. Senior engineers don't scale with ramp volume. The economic question of a complexity ramp is whether technician-level work can absorb the diagnostic load that previously required engineer-level work, because the ramp itself doesn't pause.
What changes for the rework operator
Pre-deployment, when a complex diagnostic error landed at the rework station, the operator had three options. Hold the vehicle. Escalate to a senior engineer. Wait for the senior engineer to walk over with the diagnostic information assembled from four separate systems. The third option was the only one that produced an answer; the first two produced backlog.
- Read the configuration-agnostic schematic for the vehicle family
- Mentally filter to this VIN's actual configuration
- Escalate to a senior engineer when the schematic does not resolve the fault
- Hold the vehicle; backlog accumulates at the station
- Wait for the senior engineer to assemble data from four separate systems
- Scan the VIN at the workstation
- Production Inspector returns the VIN-specific Product Twin
- Diagnostic trouble codes link to actual physical and logical components
- 2D wiring diagram and 3D twin pair, restricted to this configuration
- Senior-engineer escalation reserved for the genuinely complex residual
The point is not the model. The point is who can use it.
The expansion thesis
The structural argument holds at any plant building on the same vehicle architecture: the ECU and harness graph is identical, so the rework knowledge transfers. What changed in late 2025 is who in the customer's organization owns that question. It stopped being a single plant's IT-side decision and became a global-rework-leadership decision, because the architecture is shared and the workflow shift travels with it.
By January 2026, the operational reality at the lead plant was unambiguous: live and running through autumn, with adjacent plants on the same architecture queueing onto the same Twin.
The load-bearing word is person, not information
The default framing of a manufacturing data layer is: show the right information to the right person at the right time. That framing is right but easy to over-rotate around the information part.
What the customer's Central-European plant demonstrates is that the load-bearing word in that sentence is not information. It is person. The data layer's job is not primarily to surface information, the senior engineer was already getting the information, just slowly. The data layer's job is to make information accessible to a different class of person: the technician at the line who didn't previously have the configuration context required to use it.
Production scales with knowledge access, not with headcount. As vehicle complexity grows faster than skilled-engineer headcount, that distinction becomes the operating constraint of any modern production line. At the customer's Central-European plant, the new electric-architecture complexity is steep enough that the constraint is operationally measurable, in tender intent across plants, in expansion thesis across architectures, in what the next twelve months of the customer's group-wide production economics requires.
Program shape
| Program | New electric-architecture ramp · lead plant, Central Europe |
|---|---|
| Live since | 2025 |
| Workflow shift | 5-step escalation (hold vehicle · escalate · wait · 4-system assembly · senior-engineer answer) → VIN scan returns the Twin at the workstation |
| Class-of-person change | Technicians answer questions that previously required senior engineers |
| Onboarding next | Adjacent plant on the same architecture |
| Architecture in scope | One platform · multiple plants on shared ECU and harness graph |