TB-7386 · REV F · Technical newsheet
Design EngineeringDevice profile
MBD Meets QIF: The Standards Pair Aiming to Wire Up Manufacturing Quality
Capvidia CEO Tomasz Luniewski argues that pairing MBD with QIF preserves engineering intent end to end — enabling automated FAI, supplier data loops and AI-ready quality, provided exchanges are validated.
By Olivia Hart4 min read861 words
Features
- Capvidia CEO Tomasz Luniewski advocates pairing MBD with QIF to preserve engineering intent and traceability across CAD, manufacturing and metrology systems
- Semantic MBD enables automated inspection planning, characteristic numbering, Bills of Characteristics and FAI report population directly from the model
- QIF provides a CAD-independent structure for geometry, PMI, inspection plans and results, but adopting the standard does not alone guarantee interoperability
- Machine-readable product definition gives AI explicit engineering requirements instead of requiring inference from images, enabling root-cause analysis and predictive quality
- Recommended rollout: start with one product family, a selected supplier and a defined workflow, then validate and measure before scaling

Traditional engineering data exchange loses tolerances, feature relationships and traceability — and Capvidia argues the fix is pairing Model-Based Definition (MBD, the practice of embedding the complete product definition in the 3D CAD model) with the Quality Information Framework (QIF, an interoperability standard for quality data). Capvidia CEO Tomasz Luniewski, speaking in an interview published on the company's site, laid out both the promise and the limits of the combination.
"Traditional exchange transfers files without consistently preserving engineering meaning and traceability," Luniewski said. "Requirements are often manually extracted and re-entered, while incompatible formats can lose tolerances and feature relationships." His prescription: MBD supplies the product definition, QIF supplies the standardized framework for exchanging it across compatible systems, and the pairing connects design requirements directly with manufacturing and quality results.
What survives the conversion?
Converting native CAD data into QIF preserves geometry together with semantic product manufacturing information (PMI, the annotated tolerances and specifications attached to model features), feature definitions, datums and their associations. Metadata, identifiers and revision information must also stay available, Luniewski stressed.
The binding matters more than the bytes. A hole's tolerance must remain connected to that hole and its datum references. Shape alone is not engineering intent; intent includes how the part should function and what requirements manufacturing and inspection must satisfy. Completeness depends on three things at once — the source CAD data, the conversion software and the receiving application — so validation remains essential.
How does semantic MBD automate inspection?
Software reading semantic, machine-readable MBD can identify inspection characteristics and tolerances directly from the model. That supports inspection planning, characteristic numbering, Bills of Characteristics and population of First Article Inspection (FAI, the documented verification of a production part against requirements) reporting fields.
Measurements link back to the original requirements, cutting manual report preparation. In Luniewski's example, software identifies a hole's nominal diameter and tolerance limits, then uses the same characteristic identifier to associate its measured result. Two caveats follow: the PMI must be complete and correctly associated with geometry, and plans, results and documentation still need validation.
Where does the approach reach beyond inspection?
Structured data also feeds production engineering. Feature definitions and tolerances can inform machining decisions in compatible CAM systems. Structured weld paths and orientation data can support simulation and offline robot programming, reducing manual path reconstruction. In welding specifically, engineers can already use structured information in compatible software to develop initial trajectories and evaluate accessibility. Machining, Luniewski said, is an emerging area.
The boundary is clear, though. MBD and QIF do not generate complete production programs. Machine capabilities, tooling, fixtures, process parameters and robot kinematics are still needed, together with program validation.
Why is machine-readable definition the foundation for AI?
AI systems fed explicit engineering requirements behave differently from AI inferring them from images. "Machine-readable product definition gives AI explicit engineering requirements rather than requiring it to infer them from images," Luniewski explained. Recommendations can be linked to specific tolerances and checked against those constraints.
Connected production and inspection histories could support root-cause analysis and predictive quality. One concrete scenario: an application investigates a recurring dimensional failure in relation to a feature, its original tolerance and the relevant manufacturing operation — but only where those data are connected. Reliable applications still require validated data, appropriate models and engineering oversight.
Does adopting a standard guarantee interoperability?
No, and Luniewski said so plainly. QIF provides a common, CAD-independent structure for geometry, PMI, inspection plans and results, and it preserves relationships between characteristics and measurements across compatible systems, reducing repeated data conversion. But adopting a standard does not automatically guarantee interoperability. Applications must support the required capabilities, and exchanges must be validated.
What does this mean for supplier collaboration?
OEMs can publish validated requirements that suppliers reuse for planning, inspection and reporting, while suppliers return structured results linked to the original characteristics. Suppliers do less preparation work; OEMs receive more useful quality feedback. The aim is a closed loop: engineering requirements flow from OEM to supplier, quality results flow back to support continuous improvement.
Both sides need compatible workflows, revision control and training. Luniewski recommended pilot projects before extending the approach across the supply chain.
What stands in the way?
The principal implementation challenges are:
- semantic data quality
- software integration
- legacy workflows
- training and supplier readiness
- revision control, traceable identifiers and intellectual property protection
His rollout advice: start with one product family, a selected supplier and a defined workflow. Validate the exchange, demonstrate traceable results and measure the benefits before expanding.
Looking ahead, Luniewski sees a transition from documents to structured data, from isolated automation to connected applications, and toward more predictive quality management — with an engineering change triggering updates to related inspection plans and documentation, subject to validation and approval. "MBD defines what must be manufactured. QIF helps communicate and connect those requirements. AI can use that information to support manufacturing and inspection decisions," he said. Progress, he added, depends on reliable data, interoperability and demonstrated benefits, with human oversight remaining important.
The question manufacturers and suppliers now face is whether pilot-level traceability gains justify the integration and training investment a full QIF-based digital thread demands.
via qifstandards.org (Original)
Filed under
- model-based-definition-mbd
- quality-information-framework-qif
- cad-interoperability
- product-manufacturing-information-pmi
- manufacturing-quality
More from Olivia Hart
Show full bio
Market editor covering technology and product news in Instruments, test and measurement at Testbench Report.
97 articles
Application notes
- Verisurf 2027 Extends Model-Based Inspection to the Enterprise Layer
- Verisurf Joins DMSC to Push QIF and Open Metrology Standards
- amsight to Lead ESA FIRST! Project DIQAM on AM Qualification
- Digital Metrology Shifts Quality From Inspection to Process Control
- Cognex Pitches Simplicity as Core Differentiator in Machine Vision Market