SYWIT develops synthetic data and specially trained AI models for industrial quality inspection. The company addresses a common challenge in manufacturing: real-world data may be limited, particularly for rare defects, while conventional machine-vision systems can require repeated manual adjustments as lighting, materials, or product variants change.

SYWIT generates realistic training data by combining CAD models, shader textures, and advanced rendering techniques. Defects and environmental conditions are systematically varied to help train more robust models, improve detection accuracy, and reduce reliance on real production images. The approach supports virtual representations of components and enables defects to be created in many variations, alongside simulated influences such as lighting and surroundings.

Applications described by SYWIT include surface inspection for coatings, painted, polished, and ground surfaces; detection of manufacturing defects in processes such as injection molding, casting, and extrusion; and maintenance-related inspection of components including brake discs, brake pads, bearing surfaces, and seals. Example defect types include scratches, dents, cracks, deformation, corrosion, and wear.