After visiting the 2026 CIIF, PLC engineers are even more anxious: AI can now write programs all by itself.
01. Shock at the CIIF: From "AI Concepts" to "AI Writing Programs"
After visiting the 2026 Shanghai CIIF, Mr. Chen, who has been engaged in PLC debugging for 12 years, feels somewhat nervous.
In previous years when visiting exhibitions, AI was still confined to display boards and promotional videos, talking about "Industrial Internet", "Digital Twin", and "Predictive Maintenance". It sounded grand, but was far removed from the daily work of frontline engineers.
This year is different.
At the Siemens booth, the Eigen engineering agent is demonstrating live: simply describing the process requirements of a conveyor line in natural language can automatically generate a complete PLC project within minutes, including hardware configuration, SCL programs, and HMI screens. The safety logic is fully in place, and even the variable names comply with industry standards (Siemens China).
At the Inovance booth, the iFA Evolution platform paired with the InoQuickPro standardized framework allows users to obtain a PLC program structure compliant with the automotive industry by simply inputting vehicle models and process parameters. Over 200 standard function blocks can be called with one click. Work that used to take two weeks can now be completed within two days (Inovance Mobile).
At the Schneider Electric booth, Automation Copilot is directly embedded into the EcoStruxure engineering platform. It can read existing programs and automatically generate test cases, find syntax errors, etc. Development efficiency has improved by 30%-50%, Schneider ...
These are not just for showcasing animation effects or making conceptual prototypes, but are engineering tools that can be truly applied to actual work and generate downloadable, usable programs.
Mr. Chen stood at the booth for over twenty minutes, feeling a chill down his spine. The programming skills formed through ten years of hard work, the accumulated library of function blocks, and the summarized debugging experience seem to be depreciating at a very rapid rate in the face of these AI tools.
He is not the only one feeling this way.
02. The Three Giants Join the Fray: AI Programming Has Evolved from a Gimmick to a Practical Tool
It is widely believed that industrial AI is still in the "storytelling" stage, but in reality, it has already quietly arrived at the desks of PLC engineers.
At this year's CIIF, companies such as Siemens, Inovance, and Schneider Electric have successively launched AI programming tools for engineering projects. These are not isolated features but span the entire lifecycle of engineering projects.
Siemens: From Writing Code to Generating Entire Projects
Siemens' Eigen engineering agent has gone the furthest. It is not a standalone chatbot but is deeply embedded in TIA Portal, capable of understanding the context of the entire project—function blocks, data types, HMI screens, and the calling relationships between them, etc. It can understand it all (Siemens Di...). The two new capabilities added this year can be described as a disruptive advantage:
- First is ECAD integration. It breaks down the barriers between electrical design and software development. Once the electrical drawings are completed, AI can automatically read the component list, I/O allocation, and circuit logic, and accordingly generate the corresponding PLC hardware configuration and basic programs. Work that previously required a week of coordination between electrical and software engineers is now fully automated.
- Second is the automatic generation of standard-compliant projects. By describing equipment requirements in natural language, for example, "a conveyor line with safety doors, including frequency conversion control, emergency stop logic, fault reset, and counting functions", AI can automatically generate a complete, industry-compliant TIA Portal project. The program structure, function blocks, alarm texts, and HMI interfaces are all ready, and it can run once downloaded to the PLC (Siemens Chinese version).
Even more impressively, it can automatically generate test cases, find and fix compilation errors, and batch-replace variables. For repetitive, mechanical, and time-consuming basic programming tasks, AI can already handle most of the work.
Inovance: Standardization + AI, Taking a Differentiated Path for Domestic Substitution
Inovance's approach is more aligned with the actual conditions of Chinese factories.
Leveraging AI through the iFA Evolution fully integrated platform and the InoQuickPro standardized programming framework does not mean writing code at will; it refers to generating industry-standard, engineered programs within a short time (Inovance Mobile).
In the automotive manufacturing industry, by inputting information such as vehicle models, station numbers, and process requirements, a complete program architecture for welding/assembly production lines can be automatically generated under the InoCar template. Over 200 standard function modules can be directly reused, ensuring high program consistency and short debugging cycles. Projects that previously required senior engineers to lead can now be completed by mid-level engineers with AI.
Inovance also combines AI with virtual commissioning, using digital twins on the iFA EX platform to simulate the operation of electromechanical systems and utilizing AI to automatically generate various test conditions, identifying integration issues during the design phase. This is particularly significant for the automotive industry, which is characterized by short project cycles and high on-site commissioning costs.
Schneider Electric: Full Lifecycle Assistance and Understanding Legacy Programs
Schneider Electric's Automation Copilot has strong capabilities in handling legacy projects.
It can generate new code and also understand old programs, automatically generate functional descriptions, organize program structures, and generate test cases. Many legacy programs in factories have been used for over a decade, with no relevant documentation or personnel to explain the logical relationships. Scanning them with AI can clarify the entire architecture (Schneider ...).
This solves a major industry challenge: a large number of legacy production lines require maintenance and upgrades, which have long relied on senior engineers familiar with the programs. With AI, the onboarding difficulty for new employees is greatly reduced, and the "irreplaceability" of senior engineers is being weakened.
The Real Anxiety Is Not Unemployment, But the Rapid Depreciation of One's Experience
When it comes to AI, many people's first reaction is, "Will engineers lose their jobs because of this?"
They won't lose their jobs immediately. But a more insidious crisis than unemployment is occurring: the half-life of engineers' experience is significantly shortening.
Ten years ago, the experience accumulation curve for a PLC engineer was as follows: five years to get started, ten years to mature, and fifteen years to become an expert. Once an expert, the experience could sustain them for another ten years; the older they were, the more valuable they became.
Things are different now. Previously, it took three years to master STEP 7 V5.5; after TIA Portal was released, it took two years to fully grasp TIA Portal V15, and now TIA Portal V22 has already been iterated several times. Just as you recently perfected traditional PLC programming, AI can already generate standard programs on its own.
The interval for technology updates has shortened from over a decade to about two years. However, the advent of AI has shortened this process to the extreme.
More importantly, AI is narrowing the gap in experience.
An engineer with three years of experience, by using good AI tools, can accomplish tasks that previously required eight years of work experience. They may not necessarily understand the underlying principles or know exactly why the code is written that way, but they can complete tasks quickly with stable quality and standardized consistency.
A senior engineer with 15 years of experience, when working with traditional methods, may not even be as efficient as a younger engineer combined with AI. Much of their experience has already been tool-ized.
This is the real anxiety for engineers over 35: it is not that you will be replaced by AI, but that younger, cheaper, and more energetic young people, armed with AI, can catch up with you in a short time.
The typing speed and routines you have accumulated over ten years can be achieved by AI in seconds.
Fourth, the Change in Industry Hiring Logic is Happening Unnoticed
In addition to the technological impact, the way talent is utilized in this industry has also changed.
Ten years ago, when automotive manufacturing enterprises hired automation engineers, they valued "project experience" the most: how many welding production lines they had worked on, whether they were familiar with the VASS standard, and whether they had debugging experience, etc. The more experience, the higher the salary.
What about now?
When leading NEV (New Energy Vehicle) companies recruit engineers, they place more emphasis on "tool usage capabilities" and "rapid learning abilities". Since standard programs have been templated and basic work has been AI-ized, there is no need to write programs from scratch. You only need to quickly understand and master the process flow, and be proficient in using various tools and equipment to integrate resources and solve emerging problems.
In the past, experience was valuable; the more you did, the more valuable you were. The current efficiency premium means the faster you work, the more you are worth. Future value lies in the architecture premium; whoever can design good systems is more valuable.
The value of "skilled workers" who only know how to write programs and debug equipment according to fixed methods is decreasing at a very rapid rate. Conversely, the value of engineers who can understand processes, set standards, build systems, and solve complex problems is actually increasing.
AI does not eliminate true laborers, but rather inefficient, mechanized, and repetitive tasks.
However, the growth path of most PLC engineers over the past decade has been towards becoming "skilled workers". We have spent a lot of time practicing how to write good SCL programs, how to quickly locate a specific fault, and how to make a function block more refined.
These skills are critical, but AI is making them less important.
03. Four Ways Out for PLC Engineers
Anxiety is ineffective; the most important thing is how to respond. Based on my understanding of the industry situation, under the wave of AI, PLC engineers have four clear paths to choose from.
The first approach is to fully utilize AI technology to become a super individual.
Do not reject AI; treat it as a means to improve work efficiency. For the same project, while others take two weeks, you can complete the task in two days using AI, and use the remaining time to research more important issues and learn higher-level knowledge. AI can help you write basic code, but it will not understand the process, the site, or the true needs of the customers. Free your energy from "writing code" to "understanding problems" and "defining problems". This way, your value will increase. A veteran proficient in AI can achieve the work efficiency of three or even five traditional engineers. This is not a crisis, but an opportunity to distance yourself from your peers.
The second path is to move upwards and become a system architect.
AI can generate programs, but it cannot design system architectures. How to layer the program structure, how to define function blocks, how to plan safety logic, how to design data interfaces, and how various devices should collaborate? These are tasks that require deep thinking and a global vision, which AI cannot accomplish. Transitioning from a "program writer" to a "program system designer", and from an executor to a rule maker, is the safest upgrade path and a height that AI cannot temporarily reach.
The third path is to develop towards the process and become an expert in process automation.
PLC is just a tool; ultimately, it must serve the process. For the same automotive welding line, an engineer with welding process knowledge is more valuable than one who only knows how to use a PLC. Although AI can write programs, it does not understand weld quality, takt time optimization, or the causes of process defects. By taking root in the industry and combining automation technology with specific processes, you can become a talent who both operates technology and understands the process. This is the true moat, a barrier that AI and newcomers can hardly cross in a short time.
The fourth path is to develop towards business and become a solution provider.
Transitioning from technology to business, shifting from "doing whatever the customer asks" to "diagnosing problems for customers and proposing the best solutions". Being able to understand customer needs, determine which problems can be solved through automation, weigh the relationship between cost and benefit, and control project progress and expected goals. Such capabilities are built on extensive project experience and a deep understanding of the industry, and cannot be replaced by AI.
04. Final Words
After visiting this CIIF, my most intuitive feeling is: the wolf of industrial AI has truly arrived.
But it is not rushing in with a gaping mouth to eat all the engineers; rather, it is quietly seeping into every corner of our daily work and life, gradually changing the way we work, our value judgments, and our personal development paths.
For PLC engineers around the age of 35, this can be considered an awkward age. Their knowledge system has aged, their energy is not as vigorous as in their youth, and coupled with heavy family pressure, their speed of learning new things has slowed down.
But thinking from another perspective, isn't this a better era in any case? Good tools can help free us from heavy repetitive labor to do more meaningful and creative things.
Thirty years ago, when PLCs emerged, some said relay electricians would lose their jobs. However, the electricians who learned to use PLCs became the earliest automation engineers and ushered in the best period of their careers.
Technology will never replace human existence. Those who only know how to use outdated technology will eventually be eliminated. Instead of worrying about whether AI will replace me, it is better to recognize it, utilize it, and master it from now on.
Because in this industry, what can save you is not experience, but continuous learning ability.