The fear is understandable, but the job is changing in a different direction
When a language model can generate code, documentation and scripts in seconds, it is reasonable for engineers to wonder what remains valuable. The answer is not “nothing.” It is context.
A factory is not a programming exercise. A machine has physical constraints, safety boundaries, maintenance realities, production targets and operators. Someone still has to understand those things and translate them into a system that works reliably.
The most valuable skill is systems thinking
An automation engineer who can see the whole chain from sensor to PLC to SCADA to MES to enterprise analytics is difficult to replace. That person understands what a signal means, where it comes from, how it is controlled and how it should be consumed by higher-level systems.
AI increases the value of this perspective because modern industrial projects cross traditional boundaries.
Learn enough data engineering
SQL, Python, time-series analysis, APIs and data modelling are now useful engineering skills. You do not need to become a full-time data scientist. You need enough fluency to work with the people building analytical systems and to understand what is possible with the data available in the plant.
Do not abandon the fundamentals
Electrical engineering, control theory, instrumentation, industrial networking and commissioning still matter. The engineer who can troubleshoot a physical system and then analyse the corresponding data has a practical advantage over someone who only understands software.
Use AI as a force multiplier
Let AI handle repetitive documentation, first-pass code generation, data exploration and summarisation. Spend the saved time on architecture, testing, root-cause analysis and system design.
This is not about working less. It is about moving engineering effort toward the decisions that have higher leverage.
The career profile that is emerging
The strongest profile is neither “traditional PLC engineer” nor “generic AI engineer.” It is an engineer who can bridge OT, software and data.
That bridge is exactly where many industrial organisations still have capability gaps, and it is likely to remain valuable even as AI tools improve.
Architecture skills become more valuable
As AI tools automate fragments of coding, the engineer who can define the architecture becomes more valuable. Someone still has to decide how a PLC connects to SCADA, where the historian sits, how the data is contextualised, and which system owns each decision.
These are not tasks that become irrelevant because code generation is faster.
Learn the language of data teams
Understand tables, joins, time-series windows, features, training data, validation sets, precision, recall and model drift. You do not need a research background in machine learning to use these concepts effectively.
The point is to communicate clearly with data engineers and to recognise when a proposed solution is missing important operational context.
Become strong at commissioning and verification
AI creates more software around the plant, which means commissioning discipline becomes more valuable. A person who can test a system under normal and abnormal conditions, read logs and trace a signal from sensor to dashboard is useful across both conventional automation and AI projects.
Keep building domain depth
Industry knowledge compounds. An engineer who understands packaging, process manufacturing, utilities or pharma production has a context that general-purpose AI systems do not automatically possess. That knowledge should be captured and used to make AI systems more relevant.
The career shift is upward
The most interesting career direction is from writing every line of code toward owning the behaviour of the complete system. AI can help with implementation; engineers remain responsible for making the system make sense.
Do not chase every new tool
Model vendors, frameworks and interfaces will change faster than industrial assets. Choose a small set of durable technical foundations: Python, SQL, APIs, time-series concepts, industrial networking, documentation practices and software version control.
Build a portfolio around real engineering problems
A useful portfolio project is not “I built a chatbot.” A stronger project shows how you collected data, defined an industrial problem, measured the baseline, selected an architecture and verified the outcome. This demonstrates engineering judgement, not just tool familiarity.
Learn to discuss risk with non-AI colleagues
A strong automation engineer can explain to operations why a model may be useful, to IT why an interface needs controls and to management why a pilot needs realistic success criteria. This translation skill becomes a career advantage because industrial AI is inherently cross-functional.
Build a career around the boundary between disciplines
Factories still need engineers who can speak to maintenance, process engineering, IT and production without translating every concept through another person. AI increases the value of this boundary because a successful project crosses all of those groups.
Practical skills worth adding
Python is useful for analysis and automation. SQL is essential for manufacturing data. Git and software testing improve collaboration. Basic statistics help evaluate models. Familiarity with APIs, containers and cloud or edge concepts makes integration conversations easier.
Do not let AI weaken troubleshooting ability
There will always be days when the network is down, the diagnostic server is unavailable and the machine has a fault nobody has seen before. Engineers who understand the physical system and can reason from first principles remain valuable precisely because the digital tools are not always available.
Use AI to multiply your engineering time
Let the tool handle drafts, data preparation and repetitive analysis. Use the saved time to learn the process, improve the architecture and document what the organisation previously knew only through individual experience.
The engineer becomes an integrator of systems and knowledge
A modern automation project increasingly contains software that engineers do not write line by line themselves. The valuable task is understanding how the parts interact. A PLC exposes the state, a historian preserves the history, an MES adds production context, and an AI service identifies a pattern. Someone has to make the whole chain coherent.
Learn to challenge AI projects constructively
Useful questions include: What physical mechanism are we modelling? What happens when the data is missing? How will an operator act on this recommendation? How will we know when the model is stale? What is the rollback path?
These questions show engineering maturity without requiring the engineer to become a machine-learning researcher.
Build credibility through measurable work
Document time saved, faults detected earlier, engineering hours avoided or improvements in troubleshooting. Measurable outcomes make AI experience credible on a CV and inside an organisation.