Manufacturing SMEs face significant costs from unplanned equipment downtime but often lack access to affordable, trustworthy AI-based maintenance solutions. Existing approaches, including planned, condition-based, and corrective maintenance strategies, frequently require significant expertise, large datasets, or expensive custom implementations, creating barriers for smaller manufacturers.
Our solution is a no-code fault detection platform that enables maintenance engineers to deploy AI models without specialist data science expertise. Unlike existing competitors, the platform addresses two critical adoption challenges. First, Bayesian active learning reduces data annotation requirements by more than 95%, allowing reliable model development from as few as 10–100 labelled examples rather than thousands. The system intelligently identifies the most valuable data points for maintenance technicians to label, minimising effort and cost. Second, every prediction is accompanied by confidence intervals, enabling engineers to make informed, risk-aware decisions and improving trust in AI-assisted maintenance.
The product will be commercialised through a subscription-based software licence, supported by optional services including on-site deployment, integration with existing industrial sensors, and sensor sourcing where required. With manufacturing SMEs representing more than 99% of manufacturing companies in the EU, there is a significant opportunity to democratise access to reliable AI-based fault detection. By reducing cost, protecting data privacy, and improving user confidence, our platform enables widespread adoption of predictive maintenance technologies across the SME manufacturing sector.
