Manufacturing
Sight Machine
Sight Machine, a leader in integrated data solutions for industrial AI, helps global manufacturers monitor operations to quickly identify and resolve production issues and minimize downtime. The company’s Operator Agent solution combines live production data, agentic AI-powered recommendations—all running on the Microsoft Azure ecosystem—and physically accurate digital twins, built with OpenUSD and NVIDIA Omniverse™ technologies. By unifying production data and streamlining operations workflows, the platform empowers operators with full context and real-time visibility into production, enabling faster, more informed decisions that boost productivity, throughput, and profit margins.
Sight Machine
Simulation / Modeling / Design
NVIDIA Omniverse
Key Takeaways
As manufacturing facilities become increasingly complex and autonomous, shop floor teams often lack the real-time insights and collaboration tools needed to make effective and proactive decisions.
Plant operators from across the automotive, pharmaceutical, and process industries are frequently stuck in reactive firefighting mode. They oversee multiple machines—often without full visibility into production—making it difficult to understand how upstream and downstream issues directly impact equipment and production. They are forced to waste valuable time walking production lines and trying to coordinate via radio communications to build a full picture, determine root-cause issues, and make critical decisions.
These teams are also overwhelmed with troubleshooting production issues, managing shift changeovers, and writing reports. As a result, critical information can be lost in shift logs, spreadsheets, and shift handover conversations, preventing them from building on past knowledge and leading to production issues repeatedly resurfacing without proper resolution.
To address these challenges and needs, Sight Machine developed Operator Agent—a unified AI platform that integrates real-time data, agentic AI, and physically accurate digital twins to turn shop floor data into actionable insights that empower plant operators to accelerate decision-making and optimize operations.
Sight Machine
“By bringing real-time analysis of thousands of data points together with unlimited AI computing and visualization power and putting this solution into the hands of the people operating the plant, Sight Machine, through its collaboration with Microsoft and NVIDIA, is delivering the next generation of capability for entire systems of production.”
Jon Sobel,
CEO and Co-Founder, Sight Machine
Real-time production data is ingested into the Operator Agent platform from many sources, including Azure IoT Operations and Microsoft Fabric. The data is integrated with Operator CoPilot, an AI assistant, powered by agentic AI recommendations, to deliver recommendations to operators. These real-time data and AI-driven recommendations are integrated with physically accurate industrial facility digital twins and delivered to operators via an intuitive UI.
Sight Machine
To build the digital twins, Sight Machine uses OpenUSD, NVIDIA Omniverse technologies. and advanced LiDAR and photogrammetry technologies from NavVis. They collaborate with Kinetic Vision to quickly turn scans of production facilities into physically accurate 3D models and then integrate them into the Operator Agent platform.
By leveraging Omniverse Kit App Streaming and NVIDIA A10 GPUs on Azure, these immersive, physically accurate digital twins are made accessible on every device with a Chromium-based browser. Operator teams engage the digital twin to visualize real-time data and AI recommendations in full context. They pinpoint the exact location of issues, collaboratively problem solve, and direct resources to quickly resolve equipment failures in the real facility.
The Operator Agent platform also gives operators access to tablet-friendly workflow features that make it easy to document equipment issues, downtimes, and non-compliance events. Digitalizing and standardizing these processes helps preserve knowledge across shift changeovers. Furthermore, these production floor insights serve not only as learning opportunities for humans but also as training data for AI algorithms. Over time, these human-labeled datasets help the AI become smarter and provide better recommendations to human operators.
The integration of these technologies into a unified AI platform is already paying dividends for Sight Machine’s customers. With more reliable, system-level performance and reduced losses on every shift, teams using Operator Agent are resolving downtime in minutes instead of hours. Operators from one large manufacturer, working at a US-based plant, have reported reduced downtime and higher throughput, resulting in up to 10% improvements in line productivity and up to 15% increases in profit margins.
With AI and digital twins rapidly accelerating industrial automation and the path to more autonomous operations, Sight Machine is working to bring advanced agentic AI, simulation, and predictive capabilities to its Operator Agent platform. As the platform’s capabilities grow and become more integrated with Sight Machine’s suite of industrial AI offerings, it will empower operations teams to unlock system-wide operational intelligence. These new capabilities will equip teams to conduct important “what-if” analysis and move from reactive firefighting to proactive, continuous improvement, making every shift smarter, safer, and more productive.
To learn more about Sight Machine and their journey with OpenUSD and NVIDIA Omniverse technologies, check out their On-Demand GTC Session and NVIDIA Omniverse community livestream. If you want to learn how to build real-time digital twin applications and workflows with OpenUSD, NVIDIA Omniverse technologies, and Microsoft Azure, check out the GitHub Repo.
Start developing interoperable 3D workflows and OpenUSD applications in NVIDIA Omniverse.