All modulesExample data
Data & Intelligence
Datasets, models and recommendations from plant data
Data & AI prepares datasets, trains small models, keeps knowledge bases and gives rule-based recommendations. The IRIS assistant only reads and never changes records.
IRISData & AIPlant AMon 08:00
Work orders · CNC-04
- CorrectiveWO-0431Corrective · Spindle
- CorrectiveWO-0447Corrective · Coolant
- CorrectiveWO-0452Corrective · Spindle
- PreventiveWO-0455Preventive · Lubrication
Read live from CMMS
Work orders of CNC-04Data & AI does not copy them.Recommendation
Schedule preventive maintenance: CNC-043 corrective work orders.Rule priority 30 %
SuggestionApplied by A. Novak
ApplyApplied
How the score is computed
- Rule
- orders ÷ 10
- Corrective orders
- 3
- Score
- 0.3
CMMS
PM - CNC-04Created from the recommendation.Preventive
Data & AI reads CMMS work orders live.
What it does
- 01
Data and models
Set up sources, streams and pipelines. Experiments train on real platform data with a small algorithm set.
- 02
Knowledge bases
Documents are searched by meaning and by keyword together. Each document has its own read permissions.
- 03
Rule-based recommendations
Threshold rules propose actions and show the basis of each score. A person applies or dismisses them.
Where the data comes from
- Other IRIS modules, read live without copying their data.
- Platform events, such as IoT alarms, used for correlations.
- Uploaded documents added to a knowledge base.
Good to know
- A recommendation score is a rule priority. It is not a model prediction or a confidence level.
- Recommendations default to suggestions a person applies or dismisses. Automatic apply can be set per recommendation rule.
- Machine, process and product knowledge covers the customer's plant. IRIS product questions use a shared knowledge base.
Screens from the IRIS demo
Sample data from the IRIS demo factory. Click a screen to enlarge it.
Read more

What AI Can Really Do in Factory Operations TodaySix uses of AI that work in plants today, what each one needs, where AI is still weak, and how to pick a first use case without the autonomous factory hype.Read article · 7 min
Why Factory AI Fails Without Connected Operational DataMost factory AI pilots stall on data, not models. The five kinds of operational data AI needs, why the flow matters, and a readiness test for one line.Read article · 7 min
When AI Should Recommend and When People Should DecideThree levels of AI in plant operations (observe, advise, act), five questions that set the level for each decision, and how to write approval rules that hold.Read article · 7 min
Try Data & AI in IRIS
Start a 14-day trial on your plant's data, or open the interactive demo with sample data.