Engineering with AI: September 2026 Roundup
Updated: Sep 29
September has been our busiest month yet for AI experiments. Five new videos went up this month, all part of the same thread: using AI, Claude, ChatGPT, and the tools that connect them to real engineering software, to speed up the everyday work of building and checking plant models. Here's what we covered, most recent first...
How Far Can Claude Go in Unified Engineering? Cable Ladder Crosses (Sep 29)
Today's challenge for Claude in AVEVA Unified Engineering: cable ladder crosses, whose side rails curve round the corners. Working through our MCP bridge with the skills we have built, Claude reads the drawings and builds each cross as its own piece of equipment, stacked the way we asked. About eighteen minutes from drawings to finished stack, and the crosses are ready for clash checking.
Explore the Frontier of Engineering with Claude and Blender (Sep 18)
We step outside AVEVA entirely and test Claude against Blender's MCP server. With no plant catalogue or specs to lean on, Claude reads the drawings alone, builds the equipment, checks its own work by rendering, and finishes with a colour pass and turntable render.
Claude and Unified Engineering: PML Script Generation Is a Start (Sep 11)
Claude now works directly inside a live Unified Engineering session through the TDS AVEVA MCP Connector, so there is no macro to hand off and run ourselves. It builds the same three vessels, gives every nozzle its correct catalogue reference, and checks its own work by reading positions back out of the model. About forty minutes end to end.
AI-Assisted PML using Claude for AVEVA Unified Engineering: From 2D Drawings to 3D Equipment (Sep 6)
Claude reads a full training project drawing set, measures dimensions off the scanned GA and equipment sheets, and pulls plant coordinates from the equipment location drawing. From that, it writes one PML macro that places three real pieces of equipment at their correct plant positions when run inside Unified Engineering.
Navisworks MCP: Documents Information to the Model Using Claude AI (Sep 1)
This time we go the other way: taking information out of drawings and putting it onto the model. Claude reads a scanned equipment drawing, matches every row to a real item in the live Navisworks model, stages everything into an Excel workbook, and stops to ask us four questions before importing anything.
Taken together, these five videos trace where our AI experiments have gone this month: from AI writing a script we run ourselves, to AI working live inside our own tools, to AI reaching into a tool we don't even build for. None of it replaces the engineer. All of it removes the repetitive part of the job so we can spend more time on the part that actually needs judgement.
If any of this looks like something your team could use, get in touch. TDS Engr Solutions is happy to talk through what AI could realistically do for your engineering workflows.
