01Describe
Formulate the required live analysis in technical terms
Users describe which signals are to be synchronized, monitored, or analyzed. The task thus remains focused on the production requirements rather than on available components.
IoT Data Hub · Real-Time Analysis
Harmonize live data. Perform analyses directly within the data stream.
AI pipelines define data analyses on live streams of machine data. Instead of simply connecting graphical blocks, more complex monitoring and data harmonization tasks can be generated using natural language as explainable code.
Part of the IoT Data Hub
Diese Fähigkeit arbeitet auf derselben Konnektivität, Semantik, Historie und Governance wie alle Data-Hub-Werkzeuge.
When live logic becomes too complex for standard building blocks
AI pipelines bridge the gap between understandable business requirements and controlled, real-time analysis.
- 01
Graphical blocks are not sufficient for complex logic
Multi-level harmonization, monitoring, or pattern recognition can quickly become confusing in purely graphical models.
- 02
Individual scripts are not subject to platform operations
Code outside the data platform is difficult to test collaboratively, version control, and reuse for other data streams.
- 03
Live logic can only be verified with great effort
Without appropriate test data and a visible execution, errors often remain hidden until the system is deployed with real machine data.
- 04
Analysis results lead to a dead end
Results must be available as live data again so that other pipelines, data sinks, or notifications can respond to them.
Graphics pipelines and AI pipelines complement each other
Choosing the Right Modeling Approach
Reusable standard logic remains easy to navigate in the graphical editor. AI pipelines extend the workflow with explainable code for more complex live analyses and harmonization steps.
Request a 30-minute demo
Graphic pipelineAI Pipeline
02Check
Understanding and Reviewing Generated Code from a Technical Perspective
The generated code remains visible, can be explained, and can be checked or modified before execution. This prevents the creation of a hidden "black box" for analysis.
03Test
Verify the logic using appropriate test data
Test data helps verify behavior and output before deployment in a production environment. Adjustments remain part of the same traceable workflow.
04Operate
Carry over results as live data on the platform
The analysis being performed continuously generates new data streams. These can be stored, visualized, or used by other pipelines and actions.
What AI Pipelines Do in Production
From the technical definition to feeding the results back into ongoing data streams.
Definition of
By language
Describe complex live monitoring and data harmonization as technical requirements.
Code
Explainable
Keep generated code visible, explain it, check it, and modify it before execution.
Execution
Live
Continuously perform analytical logic on high-frequency machine data.
Results
Further usable
Provide results as new live data streams for visualization, storage, and actions.
Installation and Operation
When Are AI Pipelines the Right Approach?
Will AI pipelines replace the graphical pipeline editor?
No. Graphical pipelines remain the primary approach for reusable standard building blocks. AI pipelines complement them in cases where an analysis or harmonization would be too rigid or too complex as a graphical model.
Can the generated code be checked before it is executed?
Yes. The code remains visible, can be explained and customized, and can be tested with test data before it is run on live production data streams.
Does the language model have access to our production data?
The actual live data does not need to be transferred to the language model. The model supports code generation; data access and execution take place in a controlled manner within the platform architecture.
When are AI laptops the better choice?
AI notebooks are suitable for exploratory queries on stored data. AI pipelines are designed for continuous harmonization, monitoring, and analysis of live data streams.
30-minute product demo
Which live analysis would you like to run as a pipeline?
We'll show you how a technical specification can be turned into understandable code, a validated test, and a live data stream in production.
Product demo with a specific live data scenario