Manual Calculations That Consume Senior Engineers
Complex or multi-stage work takes days to weeks of engineering time, and any design change means restarting. Worse, the result depends on which engineer ran it. (Level 4)
We automate four categories of AutoCAD work using the .NET API, ObjectARX, and Plant 3D SDK: custom plugin development for geometry-heavy calculations, data extraction from DWG files into structured Excel and ERP systems, batch processing across thousands of drawings, and bidirectional CAD-to-SAP integration. Compiled-code performance, not LISP scripting.
A fixed sequence of commands inside one drawing: layer renaming, title block data, purge and audit runs. Fits when: the rules never change between projects. Timeline: days to 2 weeks.
Automation with its own interface, parameters and reviewable output. Fits when: inputs vary by project and several engineers must get the same result. Timeline: 6-14 weeks.
AutoCAD Plugin DevelopmentGetting information out of drawings into a form other teams can act on, across one file or a whole repository. Fits when: procurement or PM keeps asking engineers for numbers that already exist in the model. Timeline: 6-10 weeks.
CAD Data Export & ReportingCombinatorial problems, or data crossing a system boundary into SAP. Fits when: the manual procedure was already a compromise. Timeline: 3-6 months.
CAD-ERP IntegrationComplex or multi-stage work takes days to weeks of engineering time, and any design change means restarting. Worse, the result depends on which engineer ran it. (Level 4)
Quantities, material codes and element counts exist in the geometry but reach procurement by hand. Every transfer is a chance to introduce an error. (Level 3)
The drawing gets revised, the spreadsheet does not. Downstream teams act on numbers that were correct last week. (Level 3)
The same task produces different output depending on who performed it – which makes review expensive and onboarding slow. (Level 2)
We look at the task as it runs today: how often, how long, who performs it, what breaks. The output is a level assignment – which of the four approaches actually fits.
Before anything is built, we agree what “working” means in measurable terms: what the tool takes in, what it produces, and the case that proves it. You are buying an outcome with a test attached, not a number of hours.
Your CAD standards, your existing customisations, your panel catalogues and naming conventions. Nothing you already use gets replaced.
Documented source, handover sessions for your team, and support terms agreed with the engagement – including AutoCAD version compatibility as new releases ship.
We price by scope and outcome, not by hours. What moves the figure:
Which level the task belongs to. A batch extraction tool and an optimisation engine are different orders of work, and the assessment call establishes which one you need before anything is quoted.
Your CAD environment. Existing customisations, non-standard templates, and legacy plugins that the new tool has to coexist with.
Integration surface. Whether the tool stays inside AutoCAD or has to speak to an ERP, a PDM system, or a client’s data standard.
Catalogue and rule complexity. Proprietary panel systems, discipline-specific validation rules, and client naming standards all have to be encoded and tested.
Delivery timelines are predictable and we publish them: scripts and LISP routines in days to two weeks, plugins in 6-14 weeks, extraction and batch tools in 6-10 weeks, optimisation engines and ERP integrations in 3-6 months. The investment figure is set after the technical call – a published range that ignores your catalogue and integration scope would be a number neither of us could stand behind.
The assessment comes before the proposal, and it sometimes ends with "you don't need us for this."
Three reasons this domain resists it: CAD APIs have thin public documentation, so generated code compiles and then fails on the third production drawing. The engineering logic – your panel reuse rules, your ownership conventions – is not written down anywhere a model could have learned it. And geometric kernels need numerical stability built from years of edge cases, where code that drifts a millimetre per operation is worse than none.
If you run DOKA, PERI or MEVA software, we build what it doesn't cover. We are not asking you to replace anything.
Numbers you can trace are the only kind worth quoting to your management.
AutoCAD .NET API
AutoLISP
ObjectARX (C++)
C#
.NET Framework
WPF
Anything repetitive and rule-based: drawing preparation and parametric generation, data extraction into Excel or databases, batch processing across drawing sets, engineering calculations inside the CAD environment, and data synchronisation with ERP systems. What cannot be reliably automated is judgement that varies case by case without a rule you can state.
If the task is a fixed command sequence inside one drawing and the rules never change, a script or LISP routine is enough – often written in-house. If it needs parameters, a review interface, or consistent results across several engineers, it needs a plugin.
Three questions usually settle it. Does the task have inputs that change by project? Does someone outside engineering need the output? Does it cross into another system? One yes points to Level 2, two to Level 3, three to Level 4.
Scripts and LISP routines: days to two weeks. Custom plugins: 6–14 weeks. Data extraction and batch tools: 6-10 weeks. Optimisation engines and ERP integrations: 3-6 months.
AutoLISP for lightweight fixed sequences. The .NET API in C# for plugins needing custom UI, ribbons and palettes. ObjectARX in C++ for performance-critical work requiring direct geometry kernel access. Plant 3D SDK for process and piping automation.
When the workflow is under review, when the task runs only a few times a year, or when an existing product already covers it. We say so at the assessment stage rather than after a proposal.
Yes, and this is usually the first thing the assessment examines. Legacy plugins, custom templates and in-house LISP libraries define the environment the new tool has to live in – and occasionally the cheapest answer is extending what you already have.
Bring one workflow your engineers repeat. We establish which level it belongs to, what building it would involve, and whether it’s worth building. Within 48 hours you get that in writing, with a phased approach if it makes sense.
Bring one workflow your engineers repeat. In 30 minutes we'll tell you which of the four levels it belongs to, what building it would take, and whether it's worth building at all. Within 48 hours after the call: the level assignment, a technical approach, a phased timeline, and an honest answer on payback.