SAN FRANCISCO, CA, Sep 11, 2026 – Lumafield has introduced Lumafield Quality Agent, an agentic AI system that detects manufacturing defects and investigates root causes inside its Voyager platform. The Quality Agent uses a model trained on industrial X-ray CT data and compares parts against characteristics learned from known-good products rather than predefined failure modes.
Lumafield Quality Agent analyzes X-ray and CT imagery, machine vision feeds, maintenance logs and environmental sensor data. It flags deviations that may indicate previously unidentified defects and sends the finding, along with supporting evidence, to a quality engineer. Engineers can then classify the result as acceptable variation, add it to a control plan or begin containment work.
Manufacturers use visual inspection, automated optical inspection and destructive cross-sectioning before products leave the factory. These methods look for known defects or visible conditions. Quality Agent instead looks for deviations that were not defined before inspection.
Lumafield’s Cost of Quality Report found that more than 42% of manufacturers spend at least 5% of total revenue on quality-related costs. 58% estimated that at least one-quarter of their actual quality costs are not accounted for.
“Quality teams don’t need general-purpose AI retrofitted onto the factory floor; they need domain-specific intelligence built for the physical world,” said Andreas Bastian, co-founder and CTO at Lumafield. “Quality Agent is powered by a foundation model trained on hundreds of thousands of real industrial parts. By pairing our volumetric CT data expertise with agentic workflows, quality teams can now catch complex structural defects they didn’t even know to look for, trace their origin immediately, and stop escapes before they turn into costly recalls.”
The foundation model uses Lumafield’s industrial CT dataset, which contains volumetric records of manufactured products. Lumafield says this type of data is absent from commercial AI training datasets.
In one investigation, the model analyzed 1,054 lithium-ion battery cells from 10 manufacturers. It separated the cells by manufacturer and found that two brands used the same cells produced by one OEM and resold under different branding by a rewrap vendor.
“Manufacturing is a far harder problem than it is given credit for, and robots alone will not automate it,” said Bastian. “Making something is straightforward. Making safe products that perform exactly as intended is difficult, and it depends on understanding quality and the processes behind it. Automating quality and automating process control are central to automating manufacturing as a whole.”
Source: Lumafield
About Lumafield
Lumafield develops industrial X-ray computed tomography (CT) scanners and cloud-based software for non-destructive inspection and quality control in manufacturing. Its technology enables engineers to visualize internal part structures throughout product development and production. The company’s Neptune, Triton, and Voyager platforms support both lab-scale and in-line inspection, providing data-driven insights. Lumafield serves industries such as aerospace and defense, automotive, electronics, medical devices, batteries, consumer goods, and sporting equipment. Founded in 2019, the company is headquartered in Cambridge, MA, with offices in San Francisco and Los Angeles. Lumafield has received funding from investors like IVP, Spark Capital, DCVC, Kleiner Perkins, Lux Capital, Tony Fadell’s Build Collective, and Figma founder Dylan Field.
