HazardLens streamlines the machine risk assessment process by combining AI technology with proven safety methodologies aligned with industry best practices and the Machinery Regulation (EU) 2023/1230.
When you upload a photo or provide a text description of a machine, HazardLens’ multimodal AI analyzes the input to identify potential hazards. It detects relevant machine components, configurations, and risk indicators based on a vast library of real-world industrial scenarios.
From there, HazardLens assists in identifying applicable hazards and assigns risk parameters — including Severity (S), Frequency of Exposure (F), Possibility of Avoidance (P), and Probability of Occurrence (O) — in line with our risk estimation model. These values are used to support a qualitative evaluation of each risk.
Finally, the system suggests risk reduction measures based on:
Design Improvements — Reduce or eliminate hazards by changing how the machine or process functions.
Protective Measures — Use guards, interlocks, sensing devices, or other engineered solutions to prevent exposure.
Operational Controls — Provide procedures, training, warnings, or PPE to support safe use.
Users retain full control over all inputs and outputs — every assessment can be reviewed, refined, and customized to meet internal safety requirements and external regulatory obligations.
The result is a structured, standards-aligned risk assessment ready for documentation, audit, or certification — helping teams reduce manual effort, improve consistency, and accelerate compliance.
HazardLens uses state-of-the-art multimodal AI to assist with hazard identification based on visual and textual inputs. When analyzing images of machinery, the system can identify common mechanical hazards with a high degree of reliability — typically accurate in the range of 85–90% for well-lit, clearly visible, and conventional machine configurations.
However, the accuracy of AI-generated suggestions depends on the quality of the input and the complexity of the machine. Factors such as obstructed views, customized machinery, hidden hazards, or rare configurations may reduce detection accuracy. HazardLens is not a replacement for expert judgment and does not claim full automation of compliance with safety standards or the Machinery Directive.
To ensure precision and trustworthiness:
Users can manually review, adjust, or remove AI-identified hazards
Hazards can also be entered manually or described via text, with AI-assisted suggestions
All assessments are fully editable and designed to be finalized by a qualified safety professional
HazardLens is built to accelerate the risk assessment process, not replace human expertise. It offers a powerful starting point — particularly useful for identifying typical mechanical risks — while allowing full control and customization to meet regulatory and practical safety requirements.
