I like working with AI. But I do not believe everything it tells me. I find the development in engineering exciting, and at FISEKON we try out new tools early. A lot is happening in the CAD and CAE world right now.
What EPLAN and Siemens offer today
EPLAN presented the EPLAN Copilot in May 2026. Since EPLAN Platform 2027, released in September, it has been part of the software. It answers questions in natural language based on verified EPLAN knowledge, gives step by step instructions and finds the projects in which a given part is used. EPLAN has announced test runs and the generation of parts lists, connection lists and wiring lists as the next steps.
Siemens introduced the Eigen Engineering Agent for TIA Portal in April 2026 and extended it with two new capabilities in June. It reads electrical design data in XML or AML format, flags inconsistencies, adds the devices to the TIA project and generates the PLC tags from the hardware topology. It is also meant to build a project following the Siemens Automation Framework from a machine description in plain language.
The potential behind this is large. Automating recurring tasks, finding information faster, preparing designs, generating proposals. In many places AI will take work off our hands.
Can AI create a schematic?
Parts of it, yes. A tool can transfer data from one system to the next, generate lists, fill templates and find inconsistencies a person would overlook. The decisions that make a schematic, however, are taken earlier. What structure does the plant get, and how is it designated? Which cross section, which protective device, which disconnection condition? How are the safety functions built, and which performance level is reached? How does the new section fit into a plant that has been running for twenty years and has been rebuilt three times?
These questions have to be answered, and answered for, by someone who can judge them technically. A faster tool does not change that.
Faster does not automatically mean right
A generated solution can be formally correct and still not fit. Whether it is technically right and above all fits a grown plant, the customer standard and later maintenance still has to be judged by someone who understands what they are looking at.
There is also an effect we know from macros. Automation makes quality scalable, and unfortunately errors too. An assistant that builds a project in minutes writes a wrong assumption into every page just as quickly. Why we understand templates first and adopt them second is explained in EPLAN macros and templates.
And responsibility stays where it is. Whoever releases a drawing answers for the dimensioning, the protective measures and the safety functions, no matter which tool prepared it.
AI is only as good as the structure behind it
An agent that reads design data works with what is in the project. If the reference designation is inconsistent, part data is missing or functions are not cleanly separated, it carries exactly that into the next system. An assistant that searches projects only finds what is clearly designated.
So AI does not make the work on standards and structure superfluous. It makes it more important, because its result will be processed by further tools. What a designation looks like that people and software can both work with is described in Reference designation in EPLAN.
Our rule at FISEKON
We like using AI and want to keep FISEKON technically up to date. One simple rule applies. AI may take work off our hands. We do not hand over the thinking.
Anyone who wants to prepare their EPLAN projects for the next generation of tools starts with the standard and the designation. How we go about it is described under Standardisation and Article data and macros. We handle ongoing design work as EPLAN design services and, in ELCAD and WSCAD as well, as part of our electrical design.
