Prompt Engineering & Workflow Integration. AI inside the workflow, not beside it.
Layer 03, the routine. Prompts that deliver reliably, and AI built firmly into your workflow: from design in Rhino and Grasshopper to quotes and documentation. No pretty pictures nobody uses again.
- 1:1 or team
- On site around Vienna
- Online from anywhere
- from € 179 / 2 h
AI that works inside your business
A good prompt saves minutes. Built into the workflow, it saves hours every week. That is why the training does not stop at the prompt: we find the places where work repeats in your business and build AI in exactly there, as a fixed, checked step.
- Design & modellingTrigger variants, have Grasshopper scripts described, check the results.
- Office & quotesQuote and product texts, specifications, receipts prepared for the books.
- Automated workflowsAI agents that handle a step on their own, connected to the tools you use.
Three starting points, one course
The material stays the same. What changes is the project we run it through.
Designers and planners working alone
The biggest lever when there is nobody to hand routine work to.
Studios that want to introduce it properly
Not “everyone tries something”, but defined use cases, shared templates and a rule about what does not leave the building.
Architecture, industrial design, product development
Above all the part about checking. A generated number in a final thesis that nobody recalculated is an avoidable problem.
What we go through
In this order, because each module stands on the one before. Pace and depth adapt — the order does not.
- 01
What the models can and cannot do
The honest reckoning first: which tasks in a modelling workflow pay off, which get abandoned after two weeks, and how to tell the difference in advance.
Use casesLimitsCosts - 02
Concept phase
Generating variants and moods that can be used rather than merely admired. Consistent series instead of lucky hits — and how to hold a design character across twenty images.
Image seriesStyle consistencyReference library - 03
From image to geometry
The step where most people fail. How a generated image becomes usable curves, patterns and height fields that Grasshopper can genuinely work with.
Image SamplerCurve extractionReference geometry - 04
Language as a tool inside the script
Having Python and C# components described rather than writing them — and above all checking whether the result is right. Anyone who cannot check should not use it.
GhPythonC# ScriptTest cases - 05
Handing off routine work
Part lists, labelling, renaming, boilerplate for quotes and specifications. Unglamorous, and the largest real time saving.
Batch processingDocumentationSpecification texts - 06
Checking, not believing
Generative models are persuasive when they are wrong. How to check output systematically before a number reaches a fabrication file.
Check logPlausibilityFour-eyes principle - 07
Project data and confidentiality
What may be uploaded from a live project and what may not. Client data, NDAs, copyright in generated images — the questions that get expensive later.
NDAData storageUsage rights - 08
Workflow integration in your business
The modules become a fixed workflow: where AI is used, who runs it, how results are checked and handed over. If you want, I build the automation too, up to an AI agent that handles a step on its own.
Process analysisAutomationAI agents
What you need
- Prior knowledge
- Rhino fundamentals helpThe course is built on the CAD workflow. Without the Rhino connection half of it stays abstract.
- Software
- Rhino 8 and a mainstream language modelWe work with what you already use; I do not sell tools.
- Programming
- Not requiredReading and checking code is part of the course; writing it is not.
- Scope
- Set by the goalWe fix it together after the first call — at a fixed price, not as we go.
- Format
- 1:1 online · team on site · intensive blockIn person around Vienna, online from anywhere.
- Language
- German or EnglishMixed teams are no problem.
What you can do afterwards
A defined workflow
Not “let us try some AI”, but three to five specific points in the project where it demonstrably saves time.
Images that become geometry
Generated material that gets processed further in Grasshopper instead of sitting in a folder.
A rule for the studio
Clear on what may be uploaded, who checks, and how the result is documented — before someone gets it wrong.
I watched Björn work with Rhino and Grasshopper a great deal, and either he already knew the answer or he had good, creative ways of getting to one. What he has built in the way of tools is genuinely impressive. Whenever I came to him with a question he could always help, and I had the feeling he enjoyed doing it.
Is this not the same as a general AI course?
No, and that is the point. General courses are everywhere. This one is about your workflow: geometry, fabrication data, Grasshopper definitions, quotes and documentation, and about building AI into it for good.
Do you build the integration yourself?
Yes. Either your team learns it, or I build the workflow for you and hand it over with instructions. Usually it ends up a mix of both.
Do I need Rhino for this?
It helps a great deal. Half the course is about how generated material gets used in Rhino and Grasshopper. Without that connection much of it stays theoretical.
Which tools do you use?
The ones you already have. I am tied to no product and earn nothing from a recommendation. We look at what suits your workflow and what is a toy.
What if the tools change next year?
They will. That is why the emphasis is on what stays: scoping tasks properly, checking results, protecting what is confidential. That survives every version change.
Talk first, plan after
A short call to work out where you are and what you want to build. After it you get a proposal that fits both — or the honest answer that you do not need one.