Introduction

Hazard detection alone is no longer enough in today’s fast-paced industrial environments. The real challenge lies in transforming detection into actionable safety improvements. AI is revolutionizing workplace safety by not only identifying risks but also analyzing data, prioritizing hazards, and delivering actionable recommendations that drive real change. This blog explores how AI is making safety management more proactive, efficient, and effective.

The AI Process: Scanning, Analysis, Prioritization, and Actionable Recommendations

1. Scanning

AI-powered systems continuously monitor the workplace using sensors, cameras, and other IoT devices. These tools collect real-time data on potential hazards, environmental conditions, and human behaviors that could lead to incidents. SASM’s Molossus automates this process with transferring (live) data to actionable items.

2. Analysis

Once collected, AI processes vast amounts of data to identify patterns, anomalies, and potential risks. Unlike traditional safety monitoring, AI can cross-reference multiple data points, learning from past incidents to predict future risks.

3. Prioritization

Not all hazards pose the same level of threat. AI ranks risks based on severity, frequency, and likelihood, helping safety professionals focus on the most pressing issues first. This prioritization ensures that resources are allocated effectively to maximize workplace safety.

4. Actionable Recommendations

AI doesn’t just highlight risks—it provides practical, data-driven recommendations to mitigate them. Whether it’s adjusting workflow processes, implementing additional training, or enhancing protective measures, AI enables organizations to take proactive steps toward a safer workplace.

Examples of AI-Powered Workflows in Dynamic Workplaces

Automated Risk Assessments

AI streamlines risk assessments by continuously analyzing workplace conditions and generating real-time reports. This allows safety professionals to identify and address risks without waiting for manual inspections.

Bridging the Gap Between Detection and Resolution

One of the biggest challenges in workplace safety is ensuring that detected hazards lead to real-world solutions. AI bridges this gap by integrating with safety management systems to automate workflows, track interventions, and measure effectiveness. By doing so, organizations move from a reactive to a proactive safety culture.

An example: actionable items from the picture analyses

This example continues on the previous blog you can find on our website. For the length of the blog, we have only taken a partial part of the analyses. For every risk in the first blog the GPT has created actionable items and a RACI (Responsible, Accountable, Consulted and Informed) matrix that was based on the the project’s RACI and information in the safety plan, method statements and risk analyses.

step 1: follow the steps in our first blog (https://sasm.be/ehs-risk-assessment-ai-generated-risk-assessment-with-image-recognition/)

step 2: prompt your language tool with questions such as “take into account the hierarchy of controls and propose per risk 3 mitigation options for every specific risk.”

Using the hierarchy of controls, which prioritizes elimination, substitution, engineering controls, administrative controls, and personal protective equipment (PPE), here are proposed mitigation options for each specific risk:

  • Fall Hazard
    • Large gap in the floor:
      – Elimination: Repair the floor immediately to close the gap.
      – Engineering Control: Install temporary guardrails or safety nets around the gap.
      – Administrative Control: Mark the area with warning signs and restrict access.
  • Absence of guardrails around the gap:
    – Engineering Control: Install permanent guardrails around all exposed edges.
    – Administrative Control: Implement a safety protocol that requires workers to avoid the area until barriers are installed.
    – PPE: Require workers to wear harnesses if they need to work near the edge.
  • Deteriorated flooring that could give way:
    – Elimination: Replace or repair the deteriorated sections of the flooring.
    – Engineering Control: Install temporary steel plates or other supports over the compromised areas.
    – Administrative Control: Restrict access to the deteriorated areas until repairs are made

Step 3: Prompt your GPT for the actionable items and responsibilities: “make actionable items per mitigation option. The mitigation options must be made SMART, with clear expectations. Also create a RACI.”

  • Fall Hazard: Install Temporary Guardrails or Safety Nets
    Actionable Item:
    – Specific: Install temporary guardrails or safety nets around the large gap in the floor to prevent falls.
    – Measurable: Guardrails or nets must cover 100% of the exposed area by the end of the shift.
    – Achievable: Utilize available materials on-site and ensure all personnel are informed.
    – Relevant: This directly prevents falls, a critical safety issue on the site.
    – Time-bound: Completion by the end of the current workday.
    RACI:
    – Responsible: Front line supervisor
    – Accountable: Project Manager
    – Consulted: Structural Engineer, Construction Supervisor
    – Informed: All site workers

Step 4: Upload the result into your project punch list and review the results.

In the specific project, the upload was automatically and created tasks in a task manager, assigned to the persons including deadline.

Conclusion

AI is transforming workplace safety from a detection-based approach to an action-driven strategy. By scanning, analyzing, prioritizing, and recommending solutions, AI empowers safety professionals to make informed decisions and create safer working environments. Implementing a structured RACI approach ensures clear roles and responsibilities, maximizing the effectiveness of AI-driven safety improvements.

As AI adoption continues to grow, organizations that leverage its capabilities will be better equipped to prevent accidents and protect their workforce.

A Glimpse into the Future

The automation of tasks can also help in the education of the workforce, we’ll cover that item in our next blog post on how safety data sheet management can be automated and how to turn 60+ page documents into understandable two pagers that educate the workforce.

The future, however, is much closer than one might think. Automatically providing risk assessments from footage like SASM’s Molossus is already working today.

SASM specializes in bringing the future to you. We build applications and provide tools to bring AI to EHS. We streamline your processes and study where it makes sense to automate, use generative AI, and simplify processes. Do you want to leverage AI in your organisation? Contact us at ai@sasm.be