Embedding AI into Factory Workflows
Project Prometheus
This points to an industrial AI business that wins by becoming part of the factory and engineering workflow, not just another software seat. Getting deployed means mapping the customer’s CAD, PLM, ERP, simulation, and production systems, then tuning the product around safety rules, approval steps, and plant constraints. That work creates billable implementation revenue up front, and it also makes the system sticky because ripping it out would mean rebuilding process logic across many teams and tools.
-
Prometheus is built to operate across disconnected industrial software, including design tools, simulation suites, procurement systems, and production planning. A product that moves data and decisions across that stack needs hands on integration before it can deliver value, which naturally pulls revenue toward services as well as software.
-
The closest incumbents already sell into this way. Siemens combines factory software, PLM, and NVIDIA based simulation infrastructure, while Bright Machines pairs automation software with robotic cells and production workflows. In both cases, the sale is not a simple login, it is a deployment project tied to line design and operating outcomes.
-
That makes the moat practical rather than abstract. Once an AI system is wired into validation steps, manufacturing tolerances, supplier communication, and test routines, a rival has to replace not only the model but the implementation history inside the plant. That is much harder than switching between general purpose SaaS tools.
The market is heading toward lower software purity and higher deployment depth. The winners in industrial AI will look less like classic SaaS vendors and more like long term operating partners, with recurring software revenue layered on top of integration, workflow ownership, and eventually control over real production environments.