A System Integrator’s Approach to AI

Engineers use AI to speed up research, enable multilingual brainstorming and validate approaches across different platforms, while human engineers maintain full responsibility for core code development and final validation.
Use of generic AI systems for core algorithm development is avoided due to client confidentiality requirements and the critical nature of machinery safety and production reliability.
AI tools show good results in providing high-quality starting points for templates and test cases, allowing engineers to shift focus from repetitive tasks to complex edge cases and risk management activities.

The controls integration engineers at Magnum Systems use AI tools to explain a topic, generate an example equation or model, or provide suggestions for creating a code section. These AI tools can provide detailed answers, offer opportunities to get information from various sources quickly and spot patterns in technical data.

Some of our engineers request broad information that helps them understand a topic conversationally. Others ask detailed questions about specific types of hardware, software versions, tools, or how technical instructions behave in a real-world application.

Regardless of how our engineers deploy AI in these information-gathering tasks, the valuable information derived from our use of AI always leads to our users validating the answers though the details that frequently accompany the results.

How it improves our job performance

We use AI to augment our technical expertise rather than replace it. It cuts down on the time it takes to research topics, allowing us to hold initial, virtual brainstorming sessions without involving additional participants. It offers valuable second perspectives and speeds up our execution of repetitive tasks, resulting in quicker problem solving and the capacity to prioritize strategic work directly.

Our multilingual engineers use AI to absorb information and tackle challenges more efficiently by conversing in their native language while also switching languages to gather further information. Ultimately, AI enables our controls integration team to maintain a prioritized overview of essential tasks, such as project documentation, which is crucial in the systems we develop.

Using AI to validate core control code

Today, fewer than half of Magnum’s controls integration team uses AI tools for code generation. Instead, we more often use AI to perform tasks such as brainstorming HMI layouts and user flows, identifying patterns in control code and validating solutions. We also use AI to validate approaches in software across different versions or various hardware platforms.

Our engineers regard these activities as supplementary to their project work, with AI serving as an assistant. Detailed workaround architecture and code are still fully engineered, reviewed and validated by a Magnum controls integration engineer.

Based on today’s available AI tools, we believe that generic, public-access AI systems shouldn’t be used to develop algorithms for core code. A primary reason for this is that we believe any technical service provider must honor confidentiality agreements with their clients and safeguard the proprietary information they’ve been entrusted with.

Our team feels that no amount of AI can replace the need for real, experienced engineers to validate detailed logic. The stakes with our clients are too high when it comes to machinery, safety and production reliability.

In the short term, the best-case scenario we see for AI use in controls integration engineering is for it to continue advancing to a point where it can consistently provide higher-quality starting points for things like templates, test cases and verification models. This will allow our engineers to spend less time on repetitive, low-level tasks and free them up to focus on edge cases and risk management.

Legal implications for system integrators

Using AI as a tool (like a calculator, CAD software or a scripting assistant) shouldn’t require disclosure as long as the human engineer critically reviews, validates and owns the final product. We are trusted advisors to our clients, not merely contractors, and it is essential to maintain trust and transparency. That means we, as professionals, are responsible for ensuring the quality, judgment and accountability behind everything we deliver, regardless of whether AI helped speed up the process.

This position will evolve if our use of AI advances toward full code generation and essential documents like testing plans. At that point, this would become a primary discussion with our clients.

Randy Otto is senior vice president, Magnum Systems, and general manager of ECS Solutions, a Magnum Systems brand, a certified member of the Control System Integrators Association (CSIA). For more information about ECS Solutions, visit its profile on the CSIA Industrial Automation Exchange.