Today, there is a lot of talk about artificial intelligence. In your opinion, is this just a trend or a real tool that is changing the industrial automation industry?
It is an absolutely real and already working tool. If earlier automation was primarily about executing pre-defined algorithms, today it is increasingly about adaptation and forecasting. This is where AI reveals its potential. We see it not as a separate "magic pill," but as a natural development of control systems – the next step will be directed from simple automation to intelligent automation.
Can you give specific examples of where AI will bring tangible benefits in technological processes?
Of course. At EKF, we see 2 key areas. 1) Predictive analytics and maintenance. AI algorithms, analyzing data from vibration, temperature, and energy consumption sensors, can predict the failure of a pump or fan motor not in a day, but in weeks. This is a transition from maintenance "on schedule" or "on failure" to maintenance "on condition." This is a huge saving on downtime and repairs. 2) Energy consumption optimization. In building management systems (BMS) and in production, AI can analyze hundreds of indicators in real time – these are electrical parameters, weather, work schedule, tariffs – and flexibly manage load, climate, lighting, reducing energy costs by 15-25%. Our hardware solutions based on PLCs, multifunctional meters, and sensors already lay the foundation for such data collection.
What are the main challenges or barriers you see for the mass adoption of AI in industry?
There are indeed barriers. There are three of them. The most serious is the digital maturity of the enterprise. AI needs data to work – a lot of high-quality, structured data. If there is no sensor system, no historical data, there is nothing to implement. The first step is competent digitalization and the creation of a unified data environment. The second is the qualification of personnel. We need not only commissioning engineers, but also data specialists, cybersecurity specialists, capable of working at the intersection with information technology. The question of trust arises: the operator must understand why the AI made this or that decision. And the third is security. The more intelligent the system, the higher its vulnerability. The implementation of AI must go hand in hand with enhanced perimeter and data protection. This is one of our key focuses in development.
How has the current import substitution situation affected the development of the AI direction in automation in Russia?
It has become a powerful catalyst. The market has realized the critical importance of technological sovereignty. There is now a boom in demand for local, AI solutions independent of external clouds. Customers need systems that can operate autonomously, on their servers, with support within the country. This has opened a window of opportunity for Russian algorithm developers and integrators. Our task as a manufacturer is to provide them with compatible, high-quality, and safe mid-level equipment (programmable controllers, operator panels) and low-level equipment (sensors, actuators).
How will the role of traditional automation components change in the age of AI – the same PLCs? Will they become smarter?
PLCs are already evolving into hybrid computers capable of executing not only deterministic logic, but also simple local AI models for emergency response. That is, intelligence will be distributed across all levels of the system, and we, as a hardware platform manufacturer, are already planning to incorporate this computing and communication capability into our new products.
How do you see a workshop or infrastructure facility in 5-7 years from an automation point of view?
I see a self-learning autonomous system. It will not just execute a program, but continuously optimize itself for changing conditions: new batches of raw materials, equipment wear, market demand. The role of a person will shift from an operator-controller to the role of a strategist and mentor. He will set goals for the system ("maximize efficiency with such quality"), and AI will propose and implement the best scenarios. And we at EKF are already building the very technological foundation on which this future will become a reality.
If you had to formulate one main piece of advice for an industrial enterprise that is just thinking about implementing AI, where would you advise them to start?
Start not with buying software or hiring an expensive specialist. Start with an internal audit of one specific, painful process. It is better if this is a process on which money directly depends: energy overruns, a high percentage of defects, frequent unplanned shutdowns. Assemble a working group of a technologist, a workshop manager, and an IT specialist. Try to describe this process in numbers: what data you already have (even if it is in paper journals), what is missing. This audit will become your roadmap. It will show whether you need to modernize sensors first, build a digital data bus, or you can immediately try a pilot. The main thing is to move from a specific practical task, and not from an abstract desire to "be in trend."