KinetixPro

AI-powered computer vision platform for autonomous workplace safety monitoring and risk prevention

The traditional model of workplace safety enforcement is a person with a clipboard walking around a facility, checking that rules are being followed. They can monitor one area at a time, for a fraction of the working day, with no way to act on what they don't see. KinetixPro replaces that model with one that monitors everything, continuously, without anyone watching.

The problem: workplace safety enforcement hasn't scaled with facility complexity

The cost of workplace accidents is staggering in aggregate — an estimated $60 billion annually across APAC, $2 billion in Singapore alone — but the more operationally relevant number is the per-facility cost of a single serious incident. An unscheduled production line shutdown. The liability exposure from a contractor injury. The regulatory investigation following a non-compliance event. The reputational damage from a public incident. These aren't hypothetical risks; they're regular occurrences in manufacturing, warehousing, chemical processing, and logistics environments that operate at scale with large, rotating workforces.

The standard response has been CCTV — and CCTV has failed as a safety tool in one important way. It was designed to record what happened, not to prevent what's about to happen. Control rooms with human monitors exist, but the evidence on sustained human attention to passive video feeds is not encouraging: fatigue, distraction, and the sheer volume of simultaneous feeds mean that pre-incident behavioral signals go unnoticed. The insight arrives after the fact, when the only question left is liability, not prevention.

Computer vision changes the fundamental constraint. A trained AI model watching the same camera feed doesn't get tired, doesn't get distracted, and can simultaneously process every feed across an entire facility. But most computer vision safety products have addressed this at the level of individual camera detection — identifying a worker without a helmet in one frame, flagging a prohibited zone entry in another. What they've struggled to do is understand the facility as a spatial whole: tracking how a forklift is moving relative to a cluster of workers across multiple cameras, in real time, before the proximity becomes a collision.

The product: spatial intelligence, not just frame detection

KinetixPro connects to a facility's existing CCTV infrastructure — no new hardware required — and deploys a suite of AI safety models that monitor for a comprehensive range of risk scenarios: PPE compliance, area control violations, behavioural safety, housekeeping incidents, vehicle control, and facility compliance. Installation takes less than a week and is optimised for low-bandwidth environments, which matters for manufacturing and industrial sites that weren't designed with cloud-heavy software in mind.

What distinguishes KinetixPro technically is the spatial layer. By stitching together multiple camera angles, the platform builds a real-time digital twin map of the facility — tracking the positions and movements of both people and material handling equipment simultaneously across the entire site. This is architecturally different from single-camera detection. A standard computer vision system can tell you a forklift is in a frame. KinetixPro can tell you that a forklift is moving at speed toward a group of workers around a blind corner, across two cameras, and trigger an alert before either party is aware of the risk.

The operational implications extend beyond accident prevention. External vendors and contractors — who are statistically higher-risk than regular employees because they're less familiar with site-specific protocols — can be monitored against the same compliance standards as full-time workers without requiring manual supervision. Night shifts, which carry higher incident rates due to reduced oversight and worker fatigue, are monitored with the same fidelity as peak hours. The system doesn't rest when the safety manager goes home.

The compliance dimension: a regulatory pull, not just a commercial push

Workplace safety is not a discretionary spend category for enterprise customers. Food manufacturing requires stringent GMP compliance — automated cleanroom PPE and sanitation monitoring isn't a convenience, it's a regulatory requirement with real enforcement consequences. Pharmaceutical and chemical processing operate under similarly strict frameworks. Construction, warehousing, and logistics have occupational health and safety obligations that carry legal liability when violated.

This creates a buyer dynamic that is meaningfully different from typical enterprise software. A Chief Safety Officer evaluating KinetixPro isn't asking "is this nice to have?" — they're asking "what's the ROI on preventing a single serious incident and the regulatory investigation it triggers?" That's a calculation with a very short payback period. The subscription model — priced from $12,500 per year for up to 50 cameras, scaling to custom enterprise pricing — is a fraction of the annual cost of a single recordable incident at a major facility, before legal exposure is considered.

Early traction: enterprise pilots at global scale

KinetixPro's early commercial progress is a meaningful signal at the pre-scale stage. Paid pilots with STMicroelectronics, Haleon, ABB, Danieli, Bosch, and Colgate — spanning Singapore, Malaysia, Thailand, China, and Mexico — represent both the quality of the customer relationships and the geographic breadth of the opportunity. These aren't experimental engagements with small regional businesses; they're production deployments at major global manufacturers who have mature, demanding EHS functions and the sophistication to evaluate competing solutions.

The diversity of the pilot industries is also a validation of the platform's horizontal applicability. Microelectronics (STMicroelectronics), consumer health (Haleon), industrial automation (ABB), steel and metals (Danieli), and consumer goods (Bosch, Colgate) each have distinct safety environments, different PPE requirements, and different regulatory frameworks. Winning trust across that range at the pilot stage — particularly from companies with ISO-certified EHS functions — demonstrates that the product is genuinely versatile rather than narrowly optimised for a single industrial context.

Privacy by design, not as an afterthought

Enterprise deployment of computer vision in the workplace requires a credible answer to the employee privacy question. KinetixPro's approach is architecturally principled: the platform processes video frames rather than recording continuously, deletes frames immediately if no safety risk is detected, and blurs or pseudonymises all human faces in processed images. Personal data is not collected or retained as part of normal operation. The platform holds ISO 27001 and SOC 2 Type II certifications and is GDPR compliant — which matters significantly for the European operations of global manufacturers deploying the system across multiple jurisdictions.

This isn't incidental. An enterprise deploying a safety AI system at a unionised manufacturing facility in Germany faces a fundamentally different regulatory and cultural environment than one deploying at a warehouse in Singapore. The privacy-by-design architecture is what makes global deployment viable, not just technically feasible in a single market.

What would have to be true for this not to work

The honest challenges are real. Enterprise EHS software is a long sales cycle — particularly at global manufacturers where procurement, IT security, legal, and safety functions all need to sign off on a new vendor touching the CCTV network. KinetixPro's plug-and-play architecture and week-long deployment timeline reduce the implementation friction significantly, but the procurement cycle itself is what it is. Growth requires sustained enterprise sales capability alongside strong product.

The competitive landscape is also active. Intenseye, Visionify, and a range of other computer vision safety platforms are competing for the same enterprise buyers. KinetixPro's differentiator — spatial multi-camera intelligence rather than single-frame detection — is technically genuine, but needs to be argued clearly in enterprise sales cycles where buyers may not immediately appreciate the distinction. Maintaining that technical lead as the category matures will require continued investment in the spatial mapping capability.

The factors that drove our conviction: a technically differentiated architecture that addresses the highest-severity incident scenarios, demonstrated traction at enterprise scale with globally recognised customers, a privacy-first design that enables deployment in regulated markets, and a team that combined the computer vision depth required to build the product with the enterprise sales experience required to sell it. That combination is rare at the seed stage and exactly what this market requires.