In today’s manufacturing landscape, data is everywhere – from technical documents and work instructions to bills of materials and CAD models. Yet much of this valuable information remains locked away in isolated environments, disconnected and underutilized. The xLenz® platform developed by COGNI aims to change that through the smart use of AI-based edge and meta-edge tools, designed to transform a fragmented data ecosystem into integrated network of knowledge. By connecting information from documents, inspection nodes, and tracking nodes, xLenz® aims to unlock real-time insights and enable continuous process optimization right at the edge – where production happens. One of the most persistent challenges in manufacturing is the so called “tribal knowledge” – the unstructured know-how that resides in scattered sources that can only be connected by experienced workers. This includes insights, shortcuts, and problem-solving strategies developed over years on the assembly line floor. Unfortunately, much of this wisdom remains fragmented across database records, individual notes, and disconnected files. It is thus challenging for an operator to find answers in such environments, often leading to increased maintenance down-times and costs.
The xLenz® knowledge platform tackles this issue head-on by capturing, structuring and contextualizing this tribal knowledge. By creating connections between datapoints and correlating the extracted information with the formal data sources, it aims to create an evolving knowledge base accessible to everyone across the organization. The long-term vision for the xLenz® product is to create an AI-powered platform that can reliably help to cope with complex challenges related to processes, components, or assembly procedures, with precise, source-based responses derived from official technical documentation. This is made possible through a thoroughly constructed graph-based database of extracted knowledge, where entities and relationships are clearly defined, referenced and updated. The system is based on AI models to extract relevant information from text, images, and tables in a multi-modal document processing pipeline.
Another key layer of the xLenz® ecosystem is made up of the tracking and recognition nodes. The tracking nodes are used to monitor material and component flow through the facility using trackable badges that can be attached to assets. By providing real-time visibility into part status and location, these nodes contribute towards having fewer process interruptions and enable smart triggers – for example: flagging deviations, logging change of states, or notifying team members once an asset enters a pre-defined area.
Within the framework of the EdgeAI project, COGNI is working on improving the recognition nodes – intelligent devices equipped with configurable sets of sensors such as cameras and scales as well as the capability to process data closer to the edge. The nodes extract relevant features locally using visual transformers and send them to a vector database for quick and accurate matching against entries in the knowledge base. These nodes perform on-device processing for large-scale object recognition, allowing operators to quickly connect a physical component, device, or document in their hand to the corresponding digital information on the platform.
Through the fusion of AI-driven algorithms, edge and meta-edge processing, as well as graph databases COGNI is aiming to deliver a platform that makes internal company knowledge more accessible, helps reduce wasted time, manual effort and process ambiguity, thus bridging the gap between physical operations and digital intelligence.
Blog signed by: COGNI team