Intelligent safety that works with you, not instead of you.
SAiFI (pronounced “Saffy”) is a vision-based AI safety system designed to add a powerful, intelligent layer of protection across industrial environments.
Crucially, SAiFI does not replace people or existing safety measures. Instead, it acts as a force multiplier, reinforcing engineering and administrative controls while keeping human oversight firmly in place.
Enhance your existing safety measures by reinforcing them with adaptive logic and fast, reliable triggers: barriers, gates, projected warnings, and alerts.
Why SAiFI Edge
Real-time monitoring of vehicle and pedestrian movement, PPE, and site-specific hazards.
See how it works→By connecting cameras, intelligent processing, and adaptive control logic, SAiFI transforms static safety infrastructure into a living, responsive system that adapts to every shift, scenario, and site change automatically.
Cameras monitor zones, hazards, and site activity in real time.
Vision AI detects vehicles, pedestrians, PPE, and site-specific hazards across one or many zones.
Adaptive, site-specific logic applies your rules and checks scenarios to prevent unsafe transitions.
Triggers barriers, gates, projected warnings, alarms, alerts, and system actions as required.
Typical actions
No disruption to the safety systems you already have.
Integrates with your existing gates, barriers, alarms, access control, PLCs, and lighting. No need to replace the safety systems that already work.
SAiFI provides its own dedicated vision hardware, calibrated for the AI models that run on it. Consistent speed, accuracy, and predictable performance across every site.
Online, offline, or hybrid configurations to match how your site operates.
From single sites to complex, multi-zone facilities, without rearchitecting.
Our state-machine architecture ensures predictable control of pedestrian and vehicle safety zones. By monitoring vehicle entry/exit, pedestrian presence, and barrier status, SAiFI prevents unsafe transitions and maintains strict segregation in complex, high-traffic environments.
Designed for predictable behaviour
Your site does not operate the same way at 3am as it does at 3pm. SAiFI does not either.
Configure different behaviour profiles for day shift and night shift, weekday and weekend, planned maintenance windows, or seasonal changes, and schedule when each one applies. SAiFI switches between them automatically, with no operator input required.
What it gives you
Runs offline too
Scheduling runs locally on every device, so profiles continue to switch as planned even if the site is fully offline.
Choose the right level of intelligence, from simple deployments to advanced multi-zone installations.
1-2 cameras
A streamlined setup with straightforward logic, ideal for quick deployment and clear outcomes.
Best for
3-7 cameras
Enhanced workflows with state-managed logic for more responsive, adaptive environments.
Best for
8-16 cameras
Built for complex environments with multiple zones and advanced logic requirements.
Best for
Trained your way.
Reliable detection takes more than off-the-shelf AI. We work with you to define the objects, behaviours, scenarios, and edge cases that matter most, then train models using purpose-collected data from your site.
What you get
The result is a detection layer trained exactly how you want it, integrated with SAiFI state logic to provide clean, trusted inputs aligned to your zones and rules, adaptable today and future-proof as your environment evolves.
Built for environments where safety and compliance are non-negotiable.
Continuous, real-time monitoring of PPE use, exclusion zones, and high-risk site interactions, without increasing supervisory burden.
Real-time monitoring of critical processes, PPE adherence, and controlled environments to reduce risk while maintaining throughput.
Enables intelligent remote safety oversight in hazardous zones, reducing the need for personnel and electronics within high-risk areas.
Real-time visual monitoring of high-risk processes, hazardous zones, and critical behaviours, detecting unsafe conditions and process deviations as they occur.
In mixed pedestrian-vehicle spaces, SAiFI can trigger projected warnings on pedestrian detection, alerting drivers around corners and reducing blind-spot risks.
On train station platforms, SAiFI can monitor a safety boundary (for example the yellow line) and trigger automated announcements or warnings when it is crossed.
Every detection, decision, and action SAiFI takes is logged with a precise timestamp, giving you a tamper-evident record you can filter, export, and share.
Review exactly what happened, when, and why, before a near-miss becomes an incident.
Support HSE and internal safety audits with dated, searchable events and outcomes.
Share trends such as PPE adherence, zone violations, and response times over any time window.
Every detection, decision, and action is logged as a timestamped event and held securely on the device. Raw camera video is not retained. Logs sync to your cloud dashboard when connectivity is available.
SAiFI Edge protects both people and data, delivering confidence at every level.
Compliance statements reflect design intent and benchmarking described in the SAiFI overview materials; implementation may vary by deployment and configuration.
SAiFI is an active safety system, not a surveillance tool. It is built around data minimisation from the ground up, so you can deploy it with confidence under UK GDPR, the Data Protection Act, and industry-specific privacy requirements.
Live streams are processed in real time, in memory, and discarded frame by frame. SAiFI does not record or retain continuous footage.
SAiFI detects safety-relevant classes such as people, vehicles, and PPE compliance. It does not perform face recognition or identify individuals.
SAiFI is not a surveillance tool. Image capture is available as an opt-in feature for approved training data collection, always arranged with you in advance. It is never enabled silently.
Event records are kept on-device by default, retention is configurable, and you choose what is kept, for how long, and who can access it.
Designed to support lawful basis under UK GDPR, typically legitimate interests for safety-critical monitoring, and compatible with workplace surveillance guidance.
SAiFI is built to operate entirely on-site. Detection, scheduling, state logic, and event logging all run locally, which means the system keeps protecting your people even if the internet drops, a VPN fails, or the site has no connectivity at all.
Detection, decisions, actions, scheduling, and event logs all run at the edge. No cloud dependency for core safety.
Monitor and interact via on-site kiosk displays, or connect directly through a scheduled engineer visit for configuration and tuning.
Where connectivity is available, add remote dashboards, over-the-air updates, and client-facing compliance reporting. Implemented seamlessly and securely, and removed just as cleanly.
If connectivity is lost on an online system, SAiFI carries on locally and resyncs events and telemetry once it is restored, with no data gaps and no downtime.
UK-based expertise with tailored training, oversight, and ongoing support.
Fully local operation. On-site oversight and scheduled engineer visits.
On-device safety with periodic remote monitoring, health checks, and annual calibration.
Remote dashboards, over-the-air updates, and continuous optimisation; no site visit required.
Warranty
12-month warranty included as standard.
Developed where it is deployed.
SAiFI is developed and refined in the same places it is deployed, in some of the most demanding and tightly regulated operations in the world.
We work shoulder-to-shoulder with safety-critical operators across pharmaceutical, chemical, ATEX-rated beverage production, mass transit, warehouse and logistics, and construction. Every feature, every update, and every model refinement starts with a real request from the people relying on SAiFI every day.
How we stay ahead
This is why SAiFI stays at the forefront of computer vision AI for active safety, because the operations we serve will not settle for anything less.
Common questions about SAiFI Edge, active safety and AI health and safety in industry.
SAiFI Edge is the active safety model in the SAiFI product family. It is a vision-based AI system that monitors industrial environments in real time, detecting pedestrians, vehicles, PPE compliance and site-specific hazards. When a risk is identified, SAiFI Edge can trigger barriers, gates, projected warnings and alerts to prevent incidents before they occur.
SAiFI Edge is the active safety engine that detects risk and drives physical interventions: gates, barriers, projected warnings and alerts. SAiFI Aegis is a workforce-facing kiosk that turns PPE rules into visible compliance, gamification and audit evidence. Edge is what acts on a hazard, Aegis is what your team sees and engages with. Both share the same underlying SAiFI vision AI.
AI health and safety systems like SAiFI Edge use computer vision to continuously monitor workplaces without fatigue, shift gaps or blind spots. They detect unsafe conditions as they happen and trigger immediate physical responses, such as locking a gate or activating a warning, far faster than manual monitoring. This creates an active safety layer that reinforces existing controls and procedures.
Active safety refers to systems that detect and respond to risk in real time, rather than relying on passive measures like signage or manual checks. SAiFI Edge is an active safety system because it observes, detects, decides and acts; triggering physical interventions such as barriers, gates, projected warnings and alerts when hazards are identified.
SAiFI Edge is designed for sub-200-millisecond response times from detection to action, faster than human reaction time. This enables real-time interventions such as activating barriers or projecting warnings the moment a hazard is identified.
No. SAiFI Edge is designed as a force multiplier that works alongside your existing safety controls, procedures and team. It reinforces engineering and administrative controls while keeping human oversight firmly in place.
SAiFI Edge is deployed across construction, pharmaceutical, beverage production (including ATEX-rated environments), chemical manufacturing, warehouse and logistics, and mass transit environments. The system adapts to each site’s specific hazards, zones and compliance requirements.
No. SAiFI Edge is an active safety system, not a CCTV or surveillance system. Live video is processed in real time and discarded frame by frame; no continuous footage is recorded or retained, and SAiFI Edge detects classes of objects and behaviours such as people, vehicles, and PPE, rather than identifying individuals. If you choose to participate in our training programme to improve accuracy for your site, image capture can be enabled as an opt-in arrangement agreed with you in advance, with full control over what is captured, when, and for how long.
Yes, SAiFI Edge is offline-first by design. All core safety functions, including detection, decisions, actions, scheduling, and event logging, run locally on each device, so the system continues to protect your people even if the site has no internet, the VPN drops, or connectivity is unreliable. Where connectivity is available, an optional online layer adds remote dashboards, updates, and client-facing compliance reporting, and if that connection is ever interrupted SAiFI Edge simply carries on locally and resyncs when it is restored.
Yes. You can configure distinct behaviour profiles, such as one for day shift and another for night shift, then schedule when each one applies. Profiles switch automatically, with manual override available for planned maintenance or exceptional situations, and every switch is logged for audit.
SAiFI Edge keeps a tamper-evident, time-stamped record of every detection, decision, and action. On offline systems this record is available via the on-site engineer dashboard during a service visit. On online systems, you can filter, export, and share the same data remotely through your cloud dashboard, ideal for HSE audits, near-miss investigations, and client-facing compliance reports.
