End-to-end automated work safety inspections at scale
De-risking
Heavy Industries
through Autonomous Solutions
Our flagship product, InspectNet, integrates advanced robotic data acquisition with regulation-compliant data processing to produce real-time, actionable hazard alerts and comprehensive site reports.
Detection Accuracy
>95%
24/7
Monitoring
50x
Faster Inspection
80%
Cost reduction
About us
Raven Intelligent Systems™, an emerging enterprise in Switzerland and an EPFL spinoff, is at the forefront of developing automated inspection solutions tailored for industries including construction, power plants, and factories. Advanced autonomy, safety, scalability, and compliance are important characteristics of our products.
The company is currently under incubation at EPFL, finalizing its prototype and working towards its MVP. You can find the roadmap below:
The project is supported by several public and private bodies:
Services
Cutting-edge automated inspection service powered by Robotics and Artificial Intelligence
Intelligent
Data Processing
Autonomous robotic systems navigate your worksite to capture high-quality, structured data without disrupting ongoing operations.
Advanced
Data Acquisiton
Real-time Alerts &
Actionable Reporting
Collected data is processed through our compliance-grade AI pipeline to identify safety risks and operational anomalies in near real-time.
Detected issues trigger immediate alerts delivered directly to safety teams. Structured, auditable reports are generated at configurable intervals to support decision-making.
Technology
Originating as a spin-off from EPFL, Raven Intelligent Systems™ integrates robotics, artificial intelligence, and computer vision into a unified inspection and compliance platform
Ultra-compact drones & docking stations
Autonomous drones weighing under 250 gr with C0-class paired with outdoor/outdoor docking stations. The system enables repeatable inspections in GNSS-denied environments with minimal infrastructure requirements.
Swarm-enabled operations
Native multi-robot coordination allows parallel inspections, delivering up to 5× throughput compared to single-robot deployments and up to 50× improvement over manual inspection workflows.
Advanced object detection, instance segmentation, and 3D scene understanding are combined with fine-tuned vision–language models to reliably identify hazards across 2D imagery and 3D spatial reconstructions.
Compliance-grade CV & AI
Automated compliance reports & hazards
Custom large language models and agent-based pipelines generate structured alerts and compliance reports at configurable intervals, ensuring timely, auditable, and actionable outputs.
Capabilities
PPE compliance
Unprotected edges
Danger zone violation
Unstable objects
Our AI pipeline analyzes a range of hazard types, including fall-related, struck-by, caught-in, temporary-structure failure, electrical, fire, and chemical risks, and is implemented in accordance with privacy-by-design principles, including anonymization.
Missing signage
Machinary proximity
Fire risk
Temporary structures
