Lumafield Quality Agent
by Lumafield
Physical-AI quality agent that catches unanticipated manufacturing defects and traces root causes from industrial CT and sensor data
Lumafield Quality Agent is an agentic AI system that automates defect detection and root-cause analysis for manufacturers, built on a foundation model trained on industrial X-ray CT scans. It is for quality and manufacturing engineering leaders who need to catch novel defects before they become recalls, across every part rather than a sample.
Lumafield Quality Agent is an agentic AI system for manufacturing quality, announced on 3 September 2026, that automates defect detection and root-cause analysis inside Lumafield's cloud-based Voyager analysis software. It is built on what the company describes as the first large-scale foundation model trained on industrial X-ray CT data, drawn from hundreds of thousands of real manufactured parts. Rather than checking each part against a predefined list of failure modes, the agent learns the characteristics of known-good product and flags any deviation, then escalates findings to quality engineers with supporting evidence for them to classify, so it targets the novel failures that typically reach customers and trigger recalls. It can correlate CT and X-ray imagery with machine-vision feeds, maintenance logs and environmental sensor data; in one published analysis the model examined 1,054 lithium-ion cells from 10 manufacturers and found two brands reselling cells from a single OEM under different branding. The agent sits on Lumafield's hardware platform, which spans the Neptune lab CT scanner, Triton production CT, the Mars inline X-ray system and the large-format Saturn scanner launched on 10 September 2026, and Voyager integrates with PLM, QMS and MES systems, with a US-only Government Cloud option for ITAR workflows. At launch Lumafield published no pricing, general-availability date or named Quality Agent customers; demonstrations are by request. Founded in 2019, Lumafield raised a $75M Series C led by IVP in March 2025 and a $50M Silicon Valley Bank growth facility in November 2025, and competes with industrial CT incumbents Nikon Metrology, ZEISS, North Star Imaging, Waygate Technologies and Comet Yxlon.
A VP of Quality or manufacturing engineering lead at a battery, medical-device, CPG or automotive maker that already uses, or is evaluating, industrial CT inspection.
Catch defects nobody anticipated by flagging any deviation from known-good parts across every scanned unit, before they turn into recalls.
At a Glance
- Category
- Industry & Government
- Pricing
- Subscription, Contact for pricing
- Target Market
- VPs of Quality, Manufacturing Engineering Leaders, COOs, CTOs
- Deployment
- Cloud-first, Hybrid
- Founded
- 2019
- Headquarters
- San Francisco, USA
Key Features
- ✓Industrial CT foundation model
Trained on hundreds of thousands of real industrial parts so it can interpret volumetric CT data that general-purpose models cannot.
- ✓Known-good anomaly detection
Learns what good product looks like and flags any deviation, catching failure modes nobody anticipated or wrote rules for.
- ✓Multi-stream root-cause analysis
Correlates CT and X-ray imagery with machine vision, maintenance logs and environmental sensors to trace defects back to process causes.
- ✓Evidence-backed escalation
Sends flagged parts to quality engineers with supporting evidence, keeping humans in control of the final defect classification decision.
- ✓Native Voyager integration
Runs inside Lumafield's cloud Voyager software alongside porosity, GD&T inspection and trend detection with control limits that engineers already use.
- ✓PLM, QMS and MES connectivity
Voyager connects to PLM, QMS and MES systems so quality findings live in a single record across sites.
Capabilities
Use Cases
- •Battery supplier verification
Analysing 1,054 lithium-ion cells from 10 manufacturers revealed two brands reselling cells from one OEM under different labels.
- •Unknown-defect detection in production
Production lines inspecting every part flag unexpected voids, cracks, misalignment or contamination instead of relying on sampled checks.
- •Root cause of recurring defects
Quality engineers correlate defect trends with maintenance logs and environmental sensor data to find the process step causing failures.
- •Large, dense assemblies
Automotive, aerospace and data-center hardware teams scan large assemblies on Saturn and analyse the results in Voyager, where the agent runs.
Ideal For
Best For
- ✓Battery manufacturers and buyers verifying cell suppliers and internal cell quality
- ✓Medical device and consumer packaged goods teams chasing recall-risk defects
- ✓Automotive and EV production lines already running CT or inline X-ray inspection
- ✓Multi-site manufacturers that want one record of quality tied to PLM, QMS and MES
Not Ideal For
- ✗Plants that require fully on-premise or air-gapped analysis software; Voyager is delivered as a cloud service, with a US-only Government Cloud as the ITAR option
- ✗Buyers who need published pricing, a general-availability date or reference customers for the agent itself, none of which were disclosed at launch
- ✗Lines looking for an autonomous reject system with no engineer in the loop, since the agent escalates deviations for human classification
Deployment
Market Analysis
Pros
- ✓Domain-specific model built on real industrial CT data, not a retrofitted general model
- ✓Designed to catch novel defects that rule-based inspection misses
- ✓Single vendor for scanners, inline inspection and cloud analysis software
- ✓SOC 2 Type II, plus a US-only Government Cloud for ITAR-controlled data
- ✓Well-funded backer base including IVP, Kleiner Perkins and Spark Capital
Cons
- ✗No pricing, general-availability date or named customers for Quality Agent were published at launch, which limits procurement diligence
- ✗Human-in-the-loop by design: the agent flags deviations for engineers to classify rather than autonomously rejecting parts
- ✗Contrary Research notes limited presence in aerospace and medical devices, where CT incumbents have decades of experience
- ✗No documented on-premise or air-gapped Voyager deployment; the ITAR option is a managed US-only cloud
Pricing
Quality Agent
Contact for pricing
- ✓Runs inside Voyager
- ✓Demonstration on request
Neptune Standard (scanner + Voyager)
From $75,000
- ✓Industrial CT scanner
- ✓Unlimited scanning and storage
- ✓Core Voyager analysis features
Lumafield has not published pricing for Quality Agent. Its hardware is sold on subscription with Voyager software bundled: Contrary Research's 2024 profile lists the Neptune Standard package starting at $75,000, while the new Saturn scanner is offered at an undisclosed 'low upfront subscription cost'. Expect the agent to be quoted as part of a platform deal rather than as a standalone SKU.
Security & Compliance
Sources
This page was written from 8 sources, 5 on domains other than lumafield.com.
- 1.lumafield.com — quality agentvendor
- 2.lumafield.com — platformvendor
- 3.lumafield.com — introducing voyager government cloudvendor
- 4.globenewswire.com — lumafield introduces quality agent a physical ai technology
- 5.engtechnica.com — lumafield launches quality agent for ai defect detection
- 6.research.contrary.com — lumafield
- 7.prnewswire.com — lumafield secures 50 million growth capital facility from si
- 8.globenewswire.com — lumafield unveils saturn large format industrial x ray ct sc
Stay Ahead of the Curve
Weekly enterprise AI insights for technology leaders. No spam, no vendor pitches—unsubscribe anytime.
SubscribeRelated Products
Sixfold
An AI underwriter that learns your book and takes submissions through to bind-ready
Casepoint IQ
Governed agentic AI for eDiscovery, FOIA and investigations, on FedRAMP High and DoD IL6 infrastructure
Aderant Agent Center
Purpose-built AI agents for law firm back-office operations, on the Stridyn platform
DISCO Advanced Research
Agentic deep research for litigation — autonomous multi-step fact investigation with an auditable reasoning trail