Some of the most important safety signals never make it into an EHS system. A person steps into an unsafe walkway. A forklift passes too close to a pedestrian. Someone misses a PPE requirement for a few seconds. A near miss happens quickly, becomes normalised, or disappears into the pace of the shift before anyone has time to report it.
That gap sits at the centre of the Europe IMPACT session, Automated Near Miss Reporting with Computer Vision AI, co-presented with HawkVision AI. The discussion focuses on a challenge many EHS teams recognise: sites may already have cameras, but footage is often reviewed only after something has gone wrong.
Computer Vision AI changes that role. Instead of using cameras only to look backwards, organisations can use their existing CCTV feeds to identify leading indicators and potential safety risks in real time, enabling earlier intervention and follow-up. The value is not simply in seeing more events. It is in making sure the right safety signals enter a workflow where EHS teams can review, prioritise, investigate, and act.
As Subhash Sharma, Founder & CEO of HawkVision AI, explained during the session, the opportunity is to “capture non-compliance and near misses” before they become something more serious. That idea reframes near miss reporting from manual activity into a more proactive safety intelligence process.
The Real Challenge: Seeing What Workers May Not Report
Near miss reporting depends heavily on people noticing something, stopping what they are doing, remembering the details, and taking time to submit a record. Even in strong safety cultures, that does not always happen.
Some events happen too quickly. Some risks become so familiar that teams stop seeing them as reportable. Some workers may be too busy, unsure whether something counts as a near miss, or reluctant to raise an issue if it feels minor at the time.
That does not mean people do not care about safety. It means the reporting process is only capturing part of the picture.
HawkVision AI’s contribution to the session is important because it addresses that exact gap. By applying Computer Vision AI to existing CCTV footage, organisations can identify defined safety events that may otherwise go unseen, including unsafe conditions, non-compliance, and potential near misses.
When those signals are connected to Near Miss Reporting Software, they can move beyond observation and become part of a structured safety workflow.
Turning Existing Cameras Into Safety Inputs
One of the strongest points from the session is that organisations do not always need to start with new infrastructure. HawkVision AI is designed to work with existing CCTV cameras, analysing selected camera feeds in real time to identify defined safety events and potential near misses.
Those use cases may include PPE compliance, pedestrian and vehicle interaction, walkway compliance, people working under suspended loads, mobile phone use, spills, blocked access points, or other site-specific risks.
The opportunity is practical. EHS teams can start with a focused set of cameras, use cases, and thresholds, then expand once the approach proves useful. Subhash described the goal as making it easy for organisations to “try it, test it, build the business case” before moving into a broader rollout.
That makes Computer Vision AI less of a future concept and more of an operational tool that can support daily safety management.
Making Computer Vision AI Actionable Through Benchmark Gensuite
Detection alone does not create improvement. A camera can identify a potential issue, but the real value comes from what happens next.
This is where Benchmark Gensuite plays a critical role. When HawkVision AI detects a defined safety event in real time, the relevant event data and supporting evidence can flow directly into Benchmark Gensuite Concern Reporting. A concern can be created automatically, supported by key details such as location, time, use case, image evidence, and, where relevant, video context.
From there, the record enters a familiar EHS workflow. Teams can review the issue, assign ownership, investigate root causes, escalate where needed, create corrective or preventive actions, and close the loop.
That connection matters because it turns an AI-detected signal into something actionable. Instead of leaving safety insights in a camera system or separate dashboard, teams can manage them within the same platform they already use for follow-up, accountability, and reporting.
If additional actions are needed, teams can also connect findings to corrective action tracking so follow-up does not get lost after the initial review.
Avoiding Data Overload
Computer Vision AI can create more visibility, but more visibility also needs better prioritisation. If a system identifies dozens of low-level observations every day, EHS teams need a way to focus on the events that matter most.
HawkVision AI can identify defined safety events in real time, while Benchmark Gensuite enables EHS teams to manage, prioritise and act on those signals within their established safety workflows.
This prevents AI from becoming another source of noise. The goal is not to flood EHS teams with alerts. The goal is to surface the right information, with the right level of urgency, so teams can respond with context.
Building Trust Around Privacy, Culture, and Safety
Computer Vision AI also needs to fit the culture of the workplace. Workers need to understand that the technology is being used to improve safety, not to monitor individuals.
That makes privacy, transparency, and governance essential. Subhash put this simply: “Every worker wants to be safe.” That distinction matters because safety technology will only earn trust when people understand its purpose.
The session pointed to practical safeguards such as processing footage at the site, transmitting only relevant safety-event information and supporting evidence where required, and applying privacy controls such as face blurring or removal of personal identifiers.
The purpose is not to watch people. The purpose is to identify safety risks and create better opportunities for prevention.
When teams understand that Computer Vision AI is being used as a safety net, not a surveillance tool, adoption becomes easier to support.
Near misses have always carried valuable lessons, but only when organisations can see them, capture them, and act on them. HawkVision AI helps identify risks that may otherwise stay hidden. Benchmark Gensuite helps turn those signals into structured workflows, accountable follow-up, and stronger safety learning.
For organisations looking to move from reactive review to proactive risk visibility, the opportunity is clear: use the data already around you, then connect it to the workflows that help teams take action.
About HawkVision AI
HawkVision AI is on a mission to revolutionise workplace safety through real-time AI-driven visual intelligence. Its platform uses advanced computer vision and machine learning to integrate with existing CCTV systems, helping organisations detect safety hazards in real time, including near misses, PPE non-compliance, and environmental risks. HawkVision AI provides real-time computer vision detection and automated safety-event alerts, enabling relevant event information and evidence to be integrated into platforms such as Benchmark Gensuite for investigation, action and reporting.
About Benchmark Gensuite
Benchmark Gensuite® delivers AI-forward, best-in-class digital EHS, Sustainability, Quality, and Risk solutions for global enterprises. Powered by a single-version, cloud-based platform and enriched by a rapidly expanding suite of generative AI tools, Benchmark Gensuite empowers more than 4 million users across industries to proactively manage risk, ensure compliance, and accelerate performance. With over two decades of domain expertise, award-winning customer service, and a commitment to continuous innovation, Benchmark Gensuite is the trusted partner for organisations seeking both operational excellence and strategic impact.
