Where does AI currently play the biggest role in industrial automation at Siemens?

Rainer Brehm: With our many years of industrial domain expertise, we integrate AI into automation systems - across our entire portfolio and our customers' value chain. We have been doing this for many years by providing our customers with tailor-made solutions for quality assurance or process optimization, for example.

Until the 'ChatGPT' moment 2.5 years ago, it was mainly about analytical AI solutions that can be used to optimize processes based on real-time data. Senseye Predictive Maintenance, for example, uses AI to predict when machines need to be serviced before they break down. This allows us to improve maintenance efficiency by up to 55% and reduce unplanned downtime by up to 50%.

The advent of generative AI has opened up new, unimagined possibilities for the industry: Now you can ask questions in natural language to the Maintenance Copilot Senseye Predictive and get answers directly for predictive maintenance. Through such advanced generative AI functions for complex industrial use cases, we want to revolutionize how companies develop and manufacture products and operate their factories. AI plus expertise plus data is a huge opportunity for German industry.

How do AI technologies contribute to the concept of the digital factory?

AI technologies are a key driver for the evolution of the digital factory. They enable us to make intelligent use of the enormous amount of data generated in modern production environments. Through machine learning, we can not only simulate processes, but also control and optimize them with foresight. This makes the digital factory not only more efficient and flexible, but also more sustainable and resilient. We have been talking about digitalization for many years - now it is finally scaling up.

Which specific use cases do you address in mechanical and plant engineering?

One example: Siemens Industrial Copilot, a generative AI assistant, supports our customers in mechanical and plant engineering along the entire value chain in both the discrete and process industries - from design to planning, engineering, operations and services.

One specific use case, for example, is the 'NX CAD' design co-pilot, which helps engineers to automate design tasks and detect errors. This enables them to create mechanical designs faster and more efficiently. The Engineering Copilot TIA helps software developers and automation engineers to generate and document PLC code faster and to design HMI panel visualizations. By creating PLC code simply by entering it in natural language, engineering teams save time and effort and significantly reduce the likelihood of errors. We will be offering a new product version of the Engineering Copilot for the TIA Portal via the Siemens Xcelerator marketplace at the PLC trade fair in November. We are also bringing the Industrial Copilot to existing portfolios, such as COMOS, for process engineers in plant engineering. This makes engineering processes faster and simpler in the process industry too.

With the Industrial Copilot for Operations, machine operators and maintenance technicians will be able to interact with machines: They will then use knowledge from existing documentation, for example worker instructions or manuals, as well as process and sensor data via IIoT and edge devices. This applies to both discrete manufacturing and the process industry. This helps, for example, with the optimization of recipes and batches, troubleshooting in complex systems or the predictive maintenance of process components.

And with the Production Copilot Insights Hub, we already have a product on the market that even non-experts can use to quickly find and rectify the causes of production problems. This avoids production downtime despite a shortage of skilled workers.

How do your AI solutions help to make processes more resilient?

Our AI solutions help our customers to overcome many of their current challenges. Firstly, the skills shortage. Almost 8 million vacancies in the manufacturing industry will not be filled by 2030. Secondly, the increasing complexity in factories. And thirdly: increasingly fierce global competition. By implementing AI technologies across the entire value chain, we are realizing massive efficiency gains. At our appliance plant in Erlangen, for example, maintenance technicians estimate that they save an average of 25% time on reactive maintenance thanks to the use of Operations Copilot. This is not at the expense of employees. On the contrary: with AI, we are also making our economy competitive in the long term. What I find particularly exciting is that the satisfaction of our employees who use these AI applications on a daily basis is higher than before. Our colleagues are proud of their superpowers.

What role does edge computing play in the implementation?

This is an important foundation: many companies that use edge computing in their manufacturing processes want to use AI-supported use cases for intelligent manufacturing. This has been our experience with Siemens Industrial Edge and our Industrial AI portfolio. The Industrial AI Suite from Siemens addresses a key challenge in manufacturing: the reliable operationalization of AI models on the store floor. It provides a standardized, secure infrastructure for edge deployment, data integration and continuous model updates, enabling fast and reliable AI inference.

Edge computing also plays a major role for our Industrial Copilots. This is because the AI assistant can be provided as Software-as-a-Service, on an IPC on site or via Industrial Edge.

How do you deal with ethical issues surrounding AI?

Our top priority is to make artificial intelligence suitable for industry: AI applications in an industrial environment must be particularly robust, accessible to everyone, secure and reliable in order to minimize risks. Secondly, the protection of sensitive data is of crucial importance and an integral part of our AI solutions. And finally, our AI is human-centered. We don't build AI to replace humans, but to empower them.

How does Siemens support its customers in scaling AI applications?

Quite simply, by making the applications as simple and intuitive as possible. I am particularly impressed by our AI-controlled system for visual inspections, Inspekto. It enables ready-to-use visual quality inspection without any knowledge of image processing solutions or the development of machine learning systems.

The same applies to our Siemens Industrial Copilot: it revolutionizes how people interact with machines. For the first time in the history of industry, I can communicate with machines in my native language. This lowers the inhibition threshold for using this groundbreaking technology. And finally, with our entire Industrial Edge portfolio, we have tailor-made offers for our customers to implement and scale AI applications safely and easily.

The aim now is to get AI into widespread use. For example, our automotive customer Audi has been able to increase the reliability of weld spatter inspections with AI.

Which industries are benefiting particularly strongly?

No industry will be able to function without the use of AI. We take an industry-agnostic approach to the foundations of our AI applications, as many challenges are similar across industries and we want to scale our offering as quickly as possible. At the same time, we work closely with other leading companies when it comes to finding customized solutions for specific industries or applications. For example, we are working with Eplan on seamless data interoperability and the integration of AI applications in machine engineering.

How do you integrate AI into your industrial software solutions?

The same applies to our software portfolio: our portfolio is strong and will become even more powerful with AI. This means that we are integrating AI functionalities such as our Industrial Copilot into our leading industrial software such as NX or Teamcenter X throughout. This means that AI supports development and users can quickly access detailed technical information with natural language input and efficiently manage complex design tasks. This makes product development faster and smarter.

How will Industrial AI develop at Siemens over the next few years?

We see agentic AI as the next step. Agentic AI autonomously solves complex, multi-level problems by recognizing data, acting with integrated tools and continuously learning from feedback. With this approach, we want to enable industry to optimize its processes and achieve productivity gains of over 50 %.
Siemens' industrial AI agents bring intelligent autonomy to industry - from simple question-answer interactions to fully automated, cross-domain processes. An orchestrator agent determines which agents should be triggered to solve specific problems. Although this vision is still under development, it is already becoming a reality. For example, the Engineering Copilot TIA will contain agents that can handle specific use cases such as PLC code generation and HMI visualization.

Physical AI will also shape the future of industrial AI at Siemens. Instead of only operating in digital environments, physical AI combines perception and action in the real world: thanks to AI and 3D vision, robots can independently grasp unknown objects - as with Simatic Robot Pick AI Pro, Siemens' entry into this new era. This is the basis for autonomous, adaptive solutions in logistics and production.

And with the acquisition of Altair Engineering Inc., Siemens is further expanding its leading role in the areas of simulation and industrial AI and is expanding the world's most comprehensive digital twin.