Industrial AI for anomaly detection, inspection, measurement and quality control

Developing AI that learns about the fundamentals of manufacturing, Matta says its AI models can detect defects, optimise processes and evaluate equipment health.

They empower manufacturers with automation and unlock new capabilities for machine makers.

Matta delivers this as a full plug-and-play system: hardware, factory integration, AI research, and software. Most deployments, it says, go live in hours, with cameras inspecting automatically after a short learning period. The AI learns the line like an apprentice, consolidates inspection, measurement and QC in one place, traces likely root causes, and helps teams fix problems before they become costly.

Matta was founded on pioneering research from the University of Cambridge's Institute for Manufacturing, where co-founders Douglas Brion, who completed a PhD in deep learning-enabled control, and Sebastian Pattinson, Associate Professor of Engineering, first met. Spun out in 2022, Matta is now a fast-growing team of 13 full-time staff with experience from MIT, Imperial, BBC R&D, Google X, and Microsoft.

From day one, the Matta team took a demand-pull approach: setting up expos at manufacturing tradeshows, visiting 50+ UK factories, sitting with engineers and operators, mapping real pain points, and building the product around them.

Matta's technology is currently deployed in factories making products such as waterproof coats, platinum wafers, speaker cabinets and robot arms. And the company is growing quickly, adding roughly two new factory installations each month. Matta's pipeline includes more than 300 manufacturers, from SMEs to defence primes and global drinks brands.

Cameras linked to Matta's AI and platform watch products on any manufacturing line for hours or a few days. The AI model learns what good looks like from experience, and can either work fully unsupervised (determining what is good or bad by itself) or be guided with expert human feedback to capture the knowledge of a factory's most skilled workers.

Images from the camera are captured and processed, and defects and anomalies are identified and logged in real time - often with such precision that manufacturers had been missing them in previous quality verification processes.

The system is very flexible. It can be used for checking surface finish, assemblies, measurement, counting and more. It is highly robust to changes in lighting, position, and other real-world variation because of its learning-based approach. The company can review all the data to define what constitutes a good part.

Beyond inspection, Matta creates a part-level record at every key step. Install Matta across the line and you can intercept defects upstream before assembly, cutting end-product escapes. If a warranty claim appears, you can trace the product's history and show, with images and data, whether the issue arose from a component, a process step, or misuse.

The cameras also capture barcode or serial number information, powering MES-like features where objects can be tracked between stations (like a digital route card) and bottlenecks and cycle times can be monitored in real time - all while qualifying every part in the factory after every process step.

With Matta's first AI product detecting errors, the company's next step is correction: a model that can predict the root cause behind an error, why it happened, and the corrective action needed to fix it. By partnering with machine OEMs, Matta is incorporating this AI into next-generation machines, powering AI-driven closed-loop control systems that can fix themselves on the fly and learn how to use materials they have not seen before.

In the long run, Matta wants to bridge the gap between design and production. By bringing real factory data into the tools engineers use to design products, Matta will help them build products that are scalable, reliable, and work right first time - even for entirely new and ambitious ideas.