Making industrial AI accessible to non-technical users

wenglor sensoric, a German sensor technology company, had built a powerful AI vision inspection tool for manufacturing quality control. The existing platform allowed factory operators to train machine learning models and run real-time defect detection without writing a single line of code. But the interface was unintuitive, visually inconsistent, and difficult to navigate for operators on the factory floor who needed to act fast.

The goal was to redesign the platform under the new wenglor AILab brand: cleaner, more approachable, and built around the actual workflow users follow: from dataset creation to model training to device deployment.

Designing around the AI loop

The core insight was that users don't think in features. They think in steps: capture images, annotate them, train a model, deploy it. The redesign made this "AI Loop" the central mental model of the platform, surfacing it as an onboarding flow and as the structural backbone for navigation.

Each module (Datasets, Models, Devices) was redesigned to surface only what the user needs at that step, removing visual noise and clarifying the relationship between actions and outcomes. Key improvements included a redesigned annotation interface, a model performance dashboard with accuracy metrics, and a streamlined admin area for license and team management.

A platform that feels as smart as the AI it runs

The redesigned wenglor AILab shipped as a white-label platform that wenglor could deploy across its industrial customer base. The clean, light interface brought consistency to what was previously a fragmented set of screens, and the guided onboarding flow significantly reduced the learning curve for new users setting up their first inspection model.