Solution Feature
AI pipelines define data analyses on live data streams of machine data. Instead of just connecting graphical blocks using drag-and-drop, more complex monitoring and data harmonization can also be generated simply using language as explainable code.
Request a demoAI pipelines in productive use
AI pipelines are designed for live data streams. They define monitoring, harmonization steps and analytical logic directly on running machine data instead of having to work with exports or individual scripts later.
The main difference to classic pipelines is that complex logic is not just composed of reusable graphical blocks. Instead, a user can use language to describe which live analysis or harmonization is required. This is particularly helpful when graphical blocks become too rigid or too complex for the actual task.
The platform generates code from this in suitable programming languages, executes it on high-frequency data streams and makes the results available again as live data in the platform. This makes even more sophisticated evaluations much more accessible in everyday life.
Classification
AI pipelines are designed for continuous monitoring, automated reactions and the continuous processing of machine data.
If a graphical pipeline is too simple for the required logic or if results are to flow directly back into running processes, AI Pipelines is the more suitable option. For exploratory one-off analyses and the examination of stored data, on the other hand, AI Notebooks are more suitable.
Compare directly in the workflow
Classic pipelines consist of graphical drag-and-drop blocks. AI pipelines supplement this approach with language-based code generation for more complex live analyses, monitoring and data harmonization.
Request a demo
From the simple definition to the feedback of results into current data streams.
Complex live monitoring and data harmonization can also be defined by voice instead of drag-and-drop.
The generated code remains visible, can be explained and is therefore traceable instead of a black box.
The generated analyses can be performed on high-frequency live data streams.
The results are available again as live data streams in the platform and can trigger further actions.
In addition to classic graphical pipelines, complex monitoring or data harmonization can be described directly using language. This results in code instead of modeling limited to blocks.
AI pipelines generate executable code in suitable programming languages. The code remains visible, can be explained and can be checked professionally and technically before productive use.
AI-supported test data can be generated to check the pipeline. The results of the execution are then available again as live data streams in the platform.
Fields of application
AI pipelines are designed for production-related tasks that specialist users understand directly, such as anomaly detection on live data streams, temperature monitoring, condition monitoring, pattern recognition in signals or the harmonization of heterogeneous machine values.
The live results returned can then be used in other pipelines to address data sinks, execute actions or send notifications to Microsoft Teams, for example.
Demo
Arrange a demo to get to know the platform, analysis and AI functions based on your questions.