How AI can improve your safety KPIs

Key Performance Indicators (KPIs) are measurable values used to monitor progress on specific business objectives. 

They are often described against headcount, for example, as the number of events per 100,000 people, or against working hours, for example, per one million hours worked. This allows KPIs to be compared across various organizations, different departments within the same organization, and over time. 

A survey of over 600 safety specialists found that the most measured safety KPI is reported accidents and injuries. 

This KPI is a reactive or lagging measure - it only tells you when things have already gone wrong. It’s the most common KPI because, with conventional approaches, it’s the easiest to measure. 

How can computer vision enhance proactive safety?

Artificial intelligence, particularly computer vision monitoring of CCTV, makes it practical to have more proactive, leading KPIs – measures that help you prevent accidents and injuries in the first place. 

Here are four safety KPIs that will help you move from reactive to proactive monitoring of occupational safety and health (OSH).

1. Support audits and inspection processes

The second most common KPI in the survey was scored from audits and inspections. Because audits are labor intensive, organizations tend to use them infrequently - perhaps once or twice a year. 

Inspections might be required more often but can result in a tick box exercise – a process to get out of the way on Friday afternoon before you go home. CV can monitor some items that audits and inspections would look at, but more accurately and in real-time, every day of the year. 

For example, a CV can be configured to detect obstacles in walkways or doors left open. During annual audits, data analysis from CV is used to assess the accuracy of the auditing process.

2. Optimizes near-miss reporting

Counting near misses was the next most popular form of safety metrics. Near miss schemes rely on people to identify and then report things that might have resulted in an accident or injury but didn’t. 

For example, if a pedestrian walks in front of a forklift truck (FLT), causing the driver to brake sharply and drop a load, that incident might be reported. How many near misses go unreported when no damage occurs? Most people won’t repeatedly report the same near miss. 

Effective near-miss analysis with AI solutions

Computer vision can be consistent and objective when reporting near misses with the help of machine learning algorithms. This technology enhances the accuracy of KPIs and improves workplace safety. 

Organizations can use this information to develop targeted safety training programs, improve signage, or modify routes to minimize risks, which enhances their safety management ROI.

3. Cultivates safe behaviors

Counting near misses – like the sharply braking FLT in the example above – gets you closer to a preventive approach than just counting accidents, but the near miss can still damage productivity. 

Well-written safe operating procedures outline the steps needed to achieve a task safely. Identifying critical points in these steps allows for early indicators of safety performance. 

For example, a safety procedure specifies tools that should be used, and what personal protective equipment (PPE) should be worn within designated work areas. 

Computer vision could report how often the correct PPE is being worn in the location required, helping organizations improve PPE compliance. 

QR codes linked to a job management system could also report the tools being used. This data would provide a leading KPI, indicating how often there was a measurable deviation from the procedure.

4. Evaluates training effectiveness

Some organizations track the number of employees receiving safety training as a KPI. While training is essential for OHS, attending a training course doesn’t guarantee that someone will apply what they have learned on the job. 

A CV could support a training KPI. For example, a safety manager might observe that some workers over-reach instead of moving within defined movement ranges. CV can be used to count how often workers over-reach. 

Although training is provided, the issue persists. After the training, the over-reaching reduces on the day shift but not on the night shift. 

Discussion with the workers reveals that the day supervisor supports the new practices, while the night supervisor prioritizes speed over safety. The CV provides the information you need to make the training more effective.

How can you ensure KPI relevance?

It has been said that “what gets measured gets done,” but it is also true that poorly set KPIs have unintended consequences. 

This includes pointless activities or under-reporting of incidents to achieve the right numbers without making the workplace any safer or healthier. 

The good news is that computer vision enables the collection of meaningful and accurate KPIs that will support a proactive approach to OHS and behavioral safety.

Advancing EHS strategies with Protex AI

Protex AI empowers EHS teams by delivering actionable insights about safety data, which aids in improving health and safety KPIs.

Safety events or rule breaches are recorded, tagged, and stored for review by teams, offering them evidence-based insights about the performance of safety protocols. 

It auto-generates real-time safety reporting and can automatically tag stakeholders or specific team members. The storyboard functionality also allows EHS teams to create automated email workflows, add documents, or even record commentary to brainstorm and implement corrective actions. 

‍