Prerequisite
Historised time series
Captured without gaps, timestamped from one source instead of five system clocks. Without that basis every model trains on noise.
Follow-up · Platform · iniationOS
One data foundation for your production
A typical follow-up after the audit: you have ERP, MES, maintenance, quality assurance and energy meters — and a different truth in each. iniationOS puts a shared, semantically described data layer underneath. Whether that comes first at your site is in the audit report.
Request an audit
01
The problem has a name: parallel truths
Order status lives in the ERP, confirmations in the MES, faults in the maintenance system and consumption in the energy meter. Four systems, four timestamps, four names for the same machine.
That is why every analysis costs rework, every KPI costs a meeting and every investment decision rests on instinct. Not because the systems are bad — but because nobody ever brought them onto a shared description.
That is exactly our work. We replace nothing. We describe your data points once, properly — provenance, timestamp, semantics, unit — and serve them to every system from the same source.
What iniationOS covers
A data foundation for all core processes, built as an extension of an active data foundation — instead of a sixth separate system.
- Production
- Orders, confirmations, OEE per machine and shift
- Quality
- Characteristics, evidence, traceability per batch
- Inventory
- Material and goods flow, receiving through dispatch
- Asset management
- Assets, maintenance plans, fault and repair history
- Energy
- Consumption per machine, order and unit — the basis for ISO 50001
- NIS2
- Evidence included, because provenance and timestamps are described from the start
02
Production data AI systems can actually use
“AI readiness” is not a model and not a tool. It is the question of whether your time series carry a clean timestamp, a clear origin and a described meaning.
Prerequisite
Described semantics
A data point is not called “DB12.DBW4” — it carries meaning, unit and asset context. Only then can a model, or a language model, use it correctly.
Result
Use cases that hold
Predictive maintenance, anomaly detection, quality prediction and natural-language queries on production data — on-premise, at the edge or in the cloud.
The groundwork AI agents can actually run on
Whether agents will ever take over tasks in your plant is not something anyone can promise today. What they need is already settled, though: an agent can only act if it can see the current state of the plant, and if the rules are written down instead of living in a few people’s heads. That is what iniationOS produces anyway — as a by-product of what you introduce it for. So you are not buying a bet on AI, you are buying a data foundation that happens to be exactly the ground such systems need.
Discuss your data situation- Rules are written
- If-then relationships are modelled, not passed on by word of mouth
- A current picture
- The digital twin shows today’s plant, not last quarter’s
- Ready to connect
- Coding and process agents read the same described data points — in loop or graph engineering, for instance
03
How the rollout runs
No big bang. We start where the benefit becomes measurable fastest.
1
Data audit
An inventory of your systems, controllers and data sources. Result: a map of what exists, what is missing and what is maintained twice.
2
Data model
We model your assets to ISA-95 and, where useful, to companion specifications, so the structure survives a system replacement.
3
Connectivity
Machines and systems are connected over OPC UA, Modbus, MQTT or BACnet and historised. Existing systems stay untouched.
4
Analysis
Dashboards per role and interfaces for ERP, MES and AI systems. From here on, reporting stops being a separate task.
Frequently asked about iniationOS
Does iniationOS replace our ERP or MES?
No, and that is deliberate. iniationOS sits underneath your systems, not next to them. ERP and MES keep reading their data — the same data, from one source. If you have no MES, iniationOS can take over those functions; the audit establishes which is cheaper in your case.
How long does the rollout take?
The data audit typically takes days, the first connected line is a matter of weeks rather than quarters. We cut the project following the 123DAILY method so that something usable exists after every step. We give you a firm frame after the audit.
Does our data have to go to the cloud?
No. iniationOS runs on-premise, at the edge or in the cloud — the architecture follows your requirements for data sovereignty and latency. In manufacturing our default is on-premise with edge preprocessing.
We want to use AI. Do we need iniationOS first?
You need a reliable data foundation — whether it comes from us is secondary. If your time series are already properly historised and semantically described, we will say so in the audit and you skip the step. In practice this is exactly where AI projects stall.
What does it cost?
We bill on value, not on timesheets. Scope is defined up front in clearly cut steps, so you know the result and the price before you commission anything. The audit is the basis for that.
Interested, but not ready to enquire?
An e-mail address is enough. We come back once with a short read on it — no newsletter.
How clean is your data really?
Tell us which systems you run and where analysis breaks down today. You get an honest assessment of which step pays off first.
Discuss your data