Siemens Uses AI Agents to Modernize Factory Software

Siemens and Google Cloud are using AI agents to tackle the massive challenge of modernizing legacy industrial software code.

Siemens, a global leader in industrial automation, has teamed up with Google Cloud to tackle one of the biggest challenges in enterprise technology: modernizing legacy software. The collaboration focuses on using advanced AI to update the complex code that runs factories, energy grids, and transportation networks. Instead of manual rewrites, the project employs "agentic workflows," where multiple AI agents work together to understand, document, and refactor old code. These agents are guided by a specialized knowledge graph containing decades of Siemens' domain expertise. This system allows the AI to break down the enormous task of code modernization into smaller, manageable parts, a process Siemens calls "slicing the elephant."

This partnership is significant because it addresses a pain point shared by countless large organizations. Many companies rely on decades-old software that is difficult and expensive to update, hindering innovation and posing security risks. The Siemens and Google Cloud approach offers a potential blueprint for how to systematically modernize these critical systems. By combining large language models with deep, domain-specific knowledge, they are creating a more automated and scalable solution than traditional methods. For CTOs and IT leaders, this represents a promising new strategy for reducing technical debt. For developers, it could mean shifting focus from maintaining archaic systems to building new features.

The success of this initiative could have wide-ranging implications beyond industrial automation. If AI agents can effectively and safely refactor complex industrial control software, the same principles could be applied to other sectors burdened by legacy systems, such as finance, healthcare, and government. The project highlights a shift from using AI for simple code generation to deploying it for complex, mission-critical software engineering tasks. As these agentic workflows mature, they may become a standard tool in the software development lifecycle, fundamentally changing how organizations manage and evolve their technology stacks over time.