Stuttgart, Germany – August 2026. At Hannover Messe, i-flow first presented its vision of the self-connecting factory: production environments in which new machines and systems can be integrated into existing industrial data architectures with significantly less manual engineering. With Release 3.4.2, this development now reaches the next decisive milestone. The AI capabilities previously available in beta are moving into the Stable Channel and becoming part of production use in i-flow.
At the center is a self-building Unified Namespace (UNS). The i-flow Agent supports key tasks that, in traditional integration projects, still need to be carried out manually: capturing machine data, understanding and classifying data points, building semantic models, creating mappings, and transferring data into a standardized UNS structure.
The goal is to turn a complex, individual integration process into an increasingly automated and repeatable workflow.
From Beta Concept to Production Use
The foundation was laid with the beta version presented at Hannover Messe. Since then, the approach has been tested together with customers and further developed based on practical feedback. Release 3.4.2 expands the i-flow Agent particularly in areas where industrial projects require more context and control.
Users can, for example, provide technical documentation, machine descriptions, or XML files directly. The Agent can incorporate this information into its work and derive structures for models and mappings from it. It also takes the current context within the i-flow platform into account. It can analyze and structure existing tags as well as set and validate their data types.
As a result, the Agent no longer works solely on the basis of abstract inputs. It can incorporate real information from the machine and system landscape directly into the configuration of the Unified Namespace.
The Bottleneck Comes After Machine Connectivity
Under the right conditions, establishing the technical connection to a machine can often be done quickly today. The greater effort begins afterwards. A data point alone says very little about what it means, which machine or function it belongs to, and how downstream systems should interpret it. In practice, teams therefore have to manually classify and type large volumes of machine data and map them to existing data models. Mappings and topic structures then need to be created before the data can be used consistently by MES, ERP, analytics, or AI applications.
This is exactly where the self-building Unified Namespace comes in. The i-flow Agent increasingly takes over the preparatory modeling and structuring work. Based on existing information, it generates proposals for data models, mappings, and placement within the Unified Namespace. The process therefore evolves from:
manual connection → manual analysis → manual modeling → manual mapping → manual structuring
to:
automatic discovery → automatic structuring → manual review → manual approval
Automation with Controlled Approval
Production-ready AI in industry requires more than automation alone. Changes to data models and integration structures must remain traceable and controllable. That is why humans remain deliberately part of the process.
In Ask Before Edits mode, changes are not applied immediately. Asset mappings proposed by the Agent must also be confirmed before they are applied. In addition, the context available to the Agent for a given request is traceable. This helps users understand the basis on which recommendations or changes were made. As a result, the user’s role shifts away from fully manual configuration toward expert review and approval.
“The real challenge in industrial connectivity is no longer just establishing the technical connection to a machine. The effort lies in understanding, structuring, and modeling the data. That is exactly where the i-flow Agent comes in. With Release 3.4.2, we are moving this approach from beta into production use — with the goal of turning an integration project that takes weeks into an increasingly automated and repeatable process.” Timo Vormweg (CTO, i-flow GmbH)
A Key Step Toward the Self-Connecting Factory
For i-flow, the self-building Unified Namespace is not an isolated AI feature. It is part of a broader evolution in industrial data architectures. In a self-connecting factory, every new machine no longer triggers a new integration project. Systems are discovered, their data is structured, and they are placed into existing models. Humans review the result instead of configuring every individual step themselves.
This is precisely what makes Release 3.4.2 significant: AI is moving from pure assistance toward the concrete automation of industrial engineering tasks. For manufacturers, this creates the opportunity to build Unified Namespace architectures faster and extend them to additional machines, lines, and plants with significantly less manual effort.
Following the beta phase, this approach now becomes part of production use with i-flow 3.4.2.
The era of self-connecting factories has begun.
