01
Technical documents
Drawings, parts lists, data sheets and programs come from data that already exists – instead of being redrawn by hand.
- Pneumatic and wiring diagrams
- I/O, terminal and cable lists
- Control programs with simulation
AI efficiency
Your team is too good for routine.
We find the tasks that cost your company hours every week – and build AI tools that do them. Checked, traceable and designed so your team stays in control.
Illustration from Project Stack – originals confidential
01Time sinks
How much of your time goes into routine?
Pick what keeps coming up in your company and set the sliders to your numbers. Everything is calculated in your browser – nothing is stored or sent.
For you that's
Let's talk about it
2.640
hours of routine a year
As much as 1,5 full-time jobs.
What we would build for it
- Draft quotes straight from the enquiry – with your prices and your text blocks.
- Draft replies to recurring enquiries that your team only has to check.
- Drawings, lists and documents straight from CAD, Excel and descriptions.
- Reports that assemble themselves from your data and check it for plausibility.
- Translations with your glossary – technical terms are right, product names stay.
- Data from PDFs, forms and emails lands in the right system automatically.
- An assistant that searches your knowledge and backs every answer with a source.
- Checking programs that reconcile drawings, lists and programs against each other.
Pick above what costs you time.
Calculation: hours × people × 44 working weeks; one full-time job ≈ 1,760 hours a year. How much of it AI can really take over is something we'll work out honestly – on your real processes.
02Areas
Where AI can work alongside you.
02
Quotes & correspondence
An enquiry becomes a finished draft – with your prices, your text blocks and your tone.
- Quotes from enquiries
- Answers to standard questions
- Follow-ups without forgetting
03
Knowledge on demand
An assistant that knows your manuals, standards and past projects – and backs every answer with a source.
- Questions to the project archive
- Faster onboarding in the team
- Finding standards and rules
04
Reports & data
Figures from ERP, Excel and machines are merged, checked and turned into reports people actually read.
- Monthly and project reports
- Taking over data from PDFs
- Plausibility checks
05
Translation & localisation
One source, every language: websites, manuals and catalogues – with your glossary, so technical terms are right.
- Websites in many languages
- Manuals and data sheets
- Product names stay as they are
06
Checking & testing
Test programs in simulation, check drawings against rules, reconcile lists – before a mistake gets expensive.
- Simulation before commissioning
- Checks against standards and rules
- Drawing ↔ list ↔ program
03Proof
Project Stack: control engineering overnight.
A machine builder needed the complete control engineering for a new pallet stacker. The request came in the evening, 84 minutes later the first complete draft was there – and the next morning the checked package went to the engineering department.
Confidential client project, anonymised. All numbers are real, the drawings are illustrations.
TimelineReal times
Day 1
The request
3D model, sequence description, drawing, video – and one sentence: “I need a pneumatic diagram, a wiring diagram, the control cabinet and a PLC program.”
Day 1
First draft
All drawings, lists and the PLC program. Plus a simulation in which the program is already running the machine.
Day 2
Feedback
“Some symbols aren't shown correctly.” Every symbol is redrawn against the standards and the manufacturers' documents.
Day 2
Second review
“There must be no errors anywhere.” Our own checking programs now read every line, every terminal, every address.
Day 2
Ready to send
Six documents, revision 02 – ready for the engineering department.
84min
To the first draft
From the request at 8:12 pm to the complete package at 9:36 pm.
27sheets
Drawings
Pneumatics, wiring and control cabinet – to ISO 1219, IEC 60617 and EN 60204-1.
13/ 13
Simulation scenarios
Passed – including an emergency stop in every step.
0
Check findings
In the final automated review across all 27 sheets.
04How we work
Four rules we always keep.
People decide.
AI delivers drafts, experts sign off. Whatever is uncertain gets marked – not hidden.
Checkable, not a black box.
Every result is checked automatically: against rules, standards, simulations. You see what was checked – and what wasn't.
One source of truth.
Data is maintained once. Documents, lists and code are generated from it – change something, and everything changes with it.
Your data stays yours.
Before we build anything, we agree together which data each system may see. What's confidential stays confidential.
05How we start
From the first conversation to a tool in daily use.
- 01
Conversation
You tell us where things get stuck. We listen and ask questions – no slides.
- 02
Potential check
We look at your processes and tell you honestly where AI helps – and where it doesn't.
- 03
Prototype
Something that runs, early: a tool that works with your real data.
- 04
Rollout
We build it into your daily work, explain it to the team and keep developing it.
06Questions
What you might be wondering.
Will AI replace our people?
No – and that isn't the goal either. We take routine off your team so there's more time for customers, development and decisions. Responsibility and sign-off stay with you.
What happens to our data?
Before we build anything, we agree together which data may be processed and where. We treat confidential projects confidentially – that's exactly why our case study is anonymised.
Isn't this just ChatGPT?
A chat window is a start. We build tools that fit your processes: with your templates, standards and data – and with built-in checks so the results hold up.
Do we need our own AI experts?
No. We build the tools so your team can use them every day, and we show how they work.
Is it worth it for small companies too?
Especially there. Where a few people carry many tasks, every hour saved makes a difference. The calculator above gives you a first impression.
How do we start?
With a conversation about your processes. Afterwards both sides know where AI really helps – and where it doesn't.
07Contact
Which task should your team never again do by hand?
Tell us about it. We'll tell you honestly whether and how AI can help.