- Home
- AI implementation
AI implementation for industrial companies, in IT and OT
More and more companies ask the same question: how do we use AI, not only in the office but also in production? We answer it with both feet on the plant floor. We are automation engineers who build AI: we connect it to your PLC and SCADA projects, alarms, historians, manuals and maintenance history, and set it up so it is safe for operations.
Start with two or three tasks where people lose time every day, give the AI read-only access to the data those tasks need, and prove it on one line or one team before rolling out. In OT, a person stays in control of every change. We run this from assessment to roll-out, for IT and OT.
Where AI pays off in an industrial company
| Area | Examples | Who gains |
|---|---|---|
| Operations (OT) | First cause of a line stop, alarm explanations, procedures step by step, shift handover | Operators, shift leaders |
| Maintenance (OT) | Likely causes from trends, history and manuals; draft work orders; instrument health checks | Technicians, planners |
| Engineering (OT) | Read and document PLC and SCADA projects, write code to your standard, migrate legacy systems | Automation engineers |
| Office and projects (IT) | RFQs and bid comparisons, technical documentation, reports, searching procedures and specifications | Project, purchasing and quality teams |
How we implement AI
- Assess. Which tasks cost the most time, which data they need, what the risks are and what the business case is. You get a short list of use cases in a clear order.
- Connect. Read-only access to the data each use case needs: SCADA and alarm history, historian, maintenance system (CMMS), PLC projects, documents. On-premise or in EU cloud regions, agreed with IT and OT security.
- Pilot. One line or one team, with success measures agreed before the start. Operators and engineers test it in their real work.
- Roll out. Training, user rights, logging and support, then the next lines and sites.
Safe by design
- Read-only access by default: the assistant does not write to PLCs or SCADA
- New engineering work is done on development or test systems
- An engineer approves every change before it goes near the plant
- Enterprise AI services whose terms exclude training on your data by default
- Set up along the four principles for AI in OT published in December 2025 by nine cyber security agencies, including Germany's BSI and the Dutch NCSC
What we build on
We specialise in Siemens and Ignition, but we are not tied to one platform or one AI vendor. Some examples of what we use today:
- Enterprise AI models, for example Claude from Anthropic and OpenAI's models, chosen per use case
- Our own connectors for Siemens TIA Portal (V16 to V20), S7-PLCSIM Advanced, WinCC V7 and Ignition, built on the Model Context Protocol, an open standard. We build connectors to other systems the same way when a project needs them
- Local models on your own hardware where data may not leave the site
Running other PLC, SCADA, MES or IT systems? Tell us what is in your plant.
Demo on a fictional plant. All data shown is illustrative. No sound; on-screen text explains each step.
Questions we get
Where do we start?
With a short assessment: one or two workshops with operations, maintenance, engineering and IT. You get a ranked list of use cases, the data each one needs, and a pilot proposal.
Do you only work on the plant floor?
No. We implement AI in IT and OT. In industrial companies the two meet: the same assistant that explains an alarm can also draft the report, the work order or the RFQ.
Is our data used to train AI models?
We use the business offerings of Anthropic and OpenAI. Both state that business data is not used for training by default. We also agree with you exactly which data the assistant may see.
Does the EU AI Act apply to us?
It depends on the use. Under the AI Omnibus, in force since 27 July 2026, high-risk obligations apply from 2 December 2027 for the systems listed in Annex III, and from 2 August 2028 for AI built into products covered by EU product law. Assistants that explain alarms or draft documents are usually not high-risk, but AI that acts as a safety component can be. We check this for each use case; for a legal opinion, ask a lawyer.
Can it run without the cloud?
Partly. Enterprise services can run in EU regions, for example through Amazon Bedrock or Google Cloud, and some tasks can run on local models on site. We choose per use case, together with your IT and OT security.
We already have a system integrator. Why you?
One team for both: we build the AI layer and do the PLC, SCADA and E&I engineering underneath, so the AI connects safely to your control systems. On systems built by others we start with a read-only scan, without changing anything.
Sources
A new line, an upgrade, or AI for your plant?
Tell us about your project and we will set up a technical call.
2800 Mechelen, Belgium E-mail: info@simsolutions.be Enterprise no. / VAT: BE 0680.932.773