Industrial automation where the process requires it
Industrial automation is less prominent in our overall positioning, but remains a distinct capability when machinery, plant, signals and operators form part of the workflow.
Typical areas
- PLC/SCADA and sequence control
- Instrumentation, signals and data quality
- Machine monitoring, alarms and history
- Operator workflows and service access
- Integration between production data and other systems
From process data to a digital workflow
Process data can support reporting, maintenance and decision-making when signal meaning, units, timestamps and quality are documented. Any use of AI must remain outside safety-critical control and operate within clear boundaries.
Robustness and service
- Fail-safe behaviour and clear fault states
- Change control and documentation
- Separation of OT and IT based on the risk profile
- Service access and maintainable design
- Human approval of operationally critical changes
OT meets AI: clear boundaries
Control logic in machinery and plant must be deterministic, testable and safe. We therefore do not place generative AI in safety-critical control logic. The PLC and SCADA layer remains in control; AI can instead work with what that layer produces: process data, alarms, logs and documentation.
The connection between OT and digital workflows is controlled: starting with one-way data extraction, separated networks and permissions, and no write operations to the control system without a separately approved and tested scope.
Process data as a foundation for decisions
Process data is often one of the most underused sources of operational insight. Structured and timestamped, it can support maintenance planning, reporting, energy monitoring and decision support, while providing office-side AI solutions with a factual basis rather than estimates.
Common questions
Is industrial automation still a service?+
Yes. It remains a complementary capability, particularly where physical and digital workflows need to work together.
Do you use AI in PLC logic?+
AI should not replace deterministic safety or control logic. Any AI is used in clearly bounded analysis or support functions.
Can production data be used in an AI app?+
Yes, provided that signals, data quality, access, OT security and the permitted use have been clarified.