From chatbot to intelligent agent

Industrial automation is evolving rapidly, with new AI forms of technology acting as key drivers of innovation. Beckhoff recognized this early on and developed TwinCAT CoAgent, an AI-based engineering assistant as an advanced development based on TwinCAT Chat. This functionality enables significantly greater speed and efficiency in industrial development processes. Its very name illustrates its role as an active, intelligent companion that helps companies achieve their optimization goals.

TwinCAT CoAgent offers comprehensive, intelligent support in engineering. As a digital assistant, it uses generative AI models as a basis for custom-designed AI agents to accelerate development processes in PLC programming, I/O configuration, HMI design, and knowledge management, helping users achieve better results in a faster and simpler process. The technological foundation is built on the most advanced generative AI models, which Beckhoff has optimized and integrated to meet the specific requirements of industrial automation. With their exceptional ability to quickly and accurately analyze and process natural language, images as well as complex technical relationships, these AI models deliver high-quality solution proposals that significantly accelerate the engineering process.

The relevant AI agents work independently to suggest concrete actions and actively assist engineers in developing control systems. In doing so, they specifically access the Beckhoff Information System to incorporate industry- and product-specific knowledge into the solutions they propose.

Placing people at the heart of automation

TwinCAT CoAgent was developed to effectively support and simplify the demanding work of engineers and developers. Final decisions and control over the AI-generated output always remain with the specialists involved, who can review, confirm, and approve all results for execution. The AI technology handles routine tasks, suggests efficient solutions, and frees up time for more complex activities.

AI technologies deliver sustainable benefits in software development, with clear competitive advantages derived from shorter development cycles, higher-quality programming results, and increased productivity.One of the ways TwinCAT CoAgent does this is by suggesting and generating automated unit tests for PLC code. This strategy significantly improves software quality and helps identify errors early on in the development process.

The AI agents within TwinCAT CoAgent

To implement the functionalities described above, TwinCAT CoAgent includes several specialized AI agents, each with their own logic. With their help, different areas of automation can be supported and optimized as required. Initially, AI agents will be available for the following functions:

• Simplified PLC programming: Creating PLC code becomes faster and more intuitive. Users can define individual requirements conveniently in natural language, after which TwinCAT CoAgent creates suitable Structured Text code. Libraries and existing project structures are automatically taken into account, ensuring that all suggestions are relevant and context-specific. Integrated code search functions and simple referencing of libraries further simplify the development process and help improve code quality.
• Streamlined I/O module configuration: Configuring I/O topologies and communication interfaces becomes much easier. Hardware modules can be created, named, and configured via the chat function. The AI provides detailed suggestions for I/O configuration, helping reduce errors and accelerate commissioning. However, users retain full control over decisions and fine-tuning at all times, ensuring both flexibility and customization.
• More innovative creation of HMI controls: TwinCAT CoAgent creates advanced user interfaces with little effort. It uses simple sketches to automatically create complete HMI screens, identify controls, flexibly adapt design themes, and suggest suitable codes for server extensions. Automatically linking PLC variables reduces the time required during the creation process. Whether the user is building comprehensive layouts or making quick adjustments, every step can be conveniently controlled via chat. The AI handles all the groundwork while leaving the user in full control to verify results and make individual adjustments as needed.
• More direct use of Beckhoff Information System: TwinCAT CoAgent has direct access to documentation in the Beckhoff Information System, enabling the AI to search it as required. This allows users to receive customized suggestions and recommendations in line with their current requirements and project context. As a result, information retrieval is streamlined, making everyday tasks much easier.

Features of TwinCAT 3 CoAgent

TwinCAT 3 CoAgent (TE1700) is the intelligent AI assistant that provides efficient support for PLC development, I/O configuration, and human-machine interface creation. Requirements can be specified in normal language and TwinCAT CoAgent suggests suitable code and configurations, taking existing project structures into account. The flexible choice of different operating modes enables tailored support depending on the task at hand. In addition, TwinCAT CoAgent offers direct access to the documentation in the Beckhoff Information System and provides context-related recommendations for accelerated, high-quality project implementation – with the following features:

• PLC code suggestions based on natural language and project context
• AI-supported I/O configuration suggestions via chat
• human-machine interface layouts and design proposals from sketches
• direct access to the Beckhoff Information System for context-related recommendations
• flexible mode selection for custom support based on the task at hand