AI for industry
AI and automation for Swiss manufacturing companies
Stoppages, changeover times, quality measurements and fault messages provide clues for better decisions in the plant. We connect approved data from controllers, MES and documentation so shift supervisors and maintenance teams can investigate them. Implementation has a separate engineering scope with requirements for plant load, network segmentation and operational safety.
Where AI makes a concrete difference in the plant
- Downtime and OEE
- “Why did line 3 stop last week, and which cause is costing us the most time?” Answers from MES and controller data instead of from spreadsheets that someone first has to copy together.
- Shift and production reports
- Compile drafts from approved shift data. Shift supervisors check sources, completeness and interpretation before a report is used.
- Knowledge from manuals and faults
- Search approved operating instructions, maintenance logs and documented faults. Answers point to sources; safety-relevant decisions remain with the responsible specialists.
- Order processing in the office
- Preparing enquiries, quotes and purchase orders, bringing together delivery dates from ERP and production. Decisions and dispatch remain with people.
Engineering with its own test and acceptance criteria
Schady Hamid, B.Eng, leads production and engineering. Planning starts with network segmentation, cycle times, plant states and operational ownership. A finance or reporting pilot does not automatically include this work.
The starting point is read-only access through OPC UA or suitable MES interfaces. Data sources, load limits and acceptance criteria are agreed with your IT/OT team. Write operations require a separate commission, risk assessment, a list of permitted operations, human approval and an audit log.
Define data flows across the entire processing chain
A local MCP server does not guarantee local data processing. The AI application, model location, logs, telemetry and other services also matter. We assess a fully local architecture or hosting you approve. For external providers, processing locations, retention and terms for data use are reviewed; permissions and data scope are restricted independently.
Limits we state up front
- No intervention in safety-related functions such as emergency stop or protective devices.
- No write access to controllers without an allowlist, human approval and an audit log.
- No production approval without agreed load limits and tests of effects on plant operation.
- No cloud requirement. The suitability of a fully local architecture is assessed against your requirements.
Frequently asked questions
Which controllers and systems does Kapitel H work with?
We assess OPC UA, interfaces to controllers such as Siemens S7, Beckhoff and B&R, and MES and control systems. Available data, approved interfaces and acceptable load limits are what matter. Feasibility and timing are agreed based on your plant.
Do we need a cloud for this?
A fully local architecture may be possible. In addition to the MCP server and model, the AI application, logs, telemetry and other services must be configured appropriately. An external platform is only an option within the data flow you approve.
What happens if the AI suggests something wrong?
Analyses must be checked against their sources; a source reference alone does not guarantee a correct answer. Write functions are a separate engineering scope with permitted operations, human approval and logging. Safety-related functions remain excluded.
Can our own people run it themselves later?
That is the goal. Maintenance and IT are trained, and the source code and documentation in German or English are handed over to you. On request, Kapitel H takes over operation and further development under an annual contract.
Let’s check whether it holds up for your systems.
We first clarify your use case, data sources, permissions and acceptance criteria. The feasibility review documents prerequisites and limits before a pilot is agreed.
Discuss your use case