An AI use-case matrix for safety professionals
It’s getting increasingly difficult to attend a safety conference, read a LinkedIn discussion thread or safety magazine article without some mention being made of artificial intelligence. In fact, from what I can see in many instances, even the LinkedIn posts and rebuttal threads are riddled with AI. At least folks can take out the AI generated icons to make it look like it’s a real person responding. But I digress.
But there’s a catch. It’s sort of like the saying around the generality of the saying that “everyone is responsible for safety” - if everyone is responsible generally than no one if responsible specifically.
In my experience to date, AI gets referenced in some very broad-brush strokes on what it “potentially” can do for safety. But as I started to probe a little on the specifics and utility of the claim, I found I wasn’t seeing as many practical, real-world examples of what people are doing with AI or how they are leveraging it for their strategic safety initiatives, systems and processes. So based on my use of selecting AI tools over the past year, I thought I’d profile just a just a few of the practical ways I’ve been using some AI applications over the past few months. And if you wonder whether I used AI to generate this article, the answer is no, with the exception of the AI Use-Case Matrix for Safety Professionals chart.
One of the areas where safety folks typically spend a lot of time and effort is in the development or use of training videos or PPT presentations. Whether purchased or developed in-house, safety training resources can end up being a very costly item on and safety department’s budget. And if you want to make your own safety training video with corporate branding and logos it can cost in the thousands. You can of course buy generic safety training videos, but they still may not have the site-specific requirements you need, based on your unique realities or regulatory environment. Enter the AI video generation tool Synthesia.
Synthesis’s marketing materials note, you can, “Turn your SOPs, slides or policies into engaging compliance videos that are easy to update and translate into 80+ languages. The Synthesia video we made would have cost 8-10k at a minimum but with a $700 annual license for unlimited videos we are able to create our safety training video on vehicle lift safety in as long as it took us to create a Word document on the key safety aspect of vehicle lift safety. We just created our script using an existing Word document of the content of the procedure, uploaded to Synthesis, selected out Avatar and accent, and in less than 5 minutes our customized video was done. Adding in our logo gave it the brand we required. And when we need to edit further with updates or clarification it’s done in a matter of minutes and it’s good to go.
One of the other areas where safety folks typically spend significant time is in the development of various types of policies and procedures, like standard operating procedures (SOP). Many people struggle to come up with a first draft of an SOP if they don’t have any historical document as a jumping off point. So, I’m typically using applications like ChatGPT and Microsoft Copilot for consideration of any1st draft of any type of SOP or PPT presentations. They are a good trigger/starting point but need expertise and customizing to make them organization specific. But they can save a lot of up-front development time. My guess is that because AI technology, while powerful, is still in its infancy with some technical issues which typically accompany first generation IT applications. So don’t ever trust an AI generated document without doing your own proof reading and evidence/fact verification. That can be embarrassing
I use Copilot to assess/trend a wide range of behavioral observations which we make using smartphones and barcode technology. Saves a ton of time on analysis and trending and provides great visuals of the data the guys can understand. Alternatively, you’ll have to sift through those observations individually or put them all into another tool like excel and run pivot tables or other data analysis on the observations. This is time consuming and tedious.
On a personal level, I was told by my family Dr. that my blood pressure was starting to trend up, so he suggested I take/track my BP over a couple of months and come back and see him with my results. Being a bit of a technology geek as a form owner of a safety software company, I took the aggregate of all readings and asked Copilot to graph the four-month trends and give a summary of high, low and average. My Dr. said it was the most detailed BP summary he's ever seen! And I’m still using this approach to this day with a monthly summary for myself as my own reference check.
So, at the end of the days it's a tool like any other. You just need to be very, very clear on what you want it to do and give clear and unambiguous instructions to get good quality results back. But I always check/verify the content and ensure I never just spit out the draft of any content. Like any tool, if you don't know how to use it, it can be a pretty useless tool. And I decided to do an AI certification course before I dove into it in depth. But you also need to be careful of rabbit holes. It can take you to some strange places. And never assume the first draft of anything generated by any AI is complete and accurate
Of course it depends on your needs. There are literally hundreds of AI tools on the market, with more and more coming every day and more undoubtedly to come. So, there’s no shortage of tools. But it’s critical you educate yourself and the tool and its application. And whether it’s open or closed source. And it’s a good idea to check with your legal or IT department on what your organization’s policy is on the use of AI tools.
Here's a list of the current top 100. But as in all things in safety and in life, do your homework and due diligence before you even commit to using any AI tool for any reason
On a practical level, a client was looking to determine how their safety management system aligned generally with 45001, as they were a COR company but wanted to ensure alignment with some of the general best practices guidelines under ISO 45001, as it was a customer requirement. We ran this scenario and were able to give them some characterization as to where ISO 45001 and COR aligned, and some typical AI tools they could consider in helping them manage their client requirements better.
Welcome to the wild and wonderful world of AI. There’s no going back.
AI Use-Case Matrix for Safety Professionals
This is an AI Use-Case Matrix mapped to key elements of a Safety Management System (SMS) based on frameworks like ISO 45001, COR, and general best practices. This format connects AI tools, practical applications, and benefits directly to SMS elements to help you identify integration points and demonstrate potential value.
|
Safety Management System Element |
AI Use Case |
Tools/Apps |
Practical Benefit |
Supports Compliance With |
|
Leadership & Commitment |
Safety dashboards & AI reporting |
Power BI + GPT, Enablon |
Real-time safety KPIs, trends to leadership |
ISO 45001 Clause 5, COR Element 1 |
|
Hazard Identification & Risk Assessment |
Computer vision for detecting hazards |
Intenseye, Protex AI, Voxel AI |
Proactive identification of unsafe acts/conditions |
ISO 45001 Clause 6.1.2, COR Elements 2, 6 |
|
Planning |
Predictive analytics to prioritize risks |
Benchmark ESG, Avetta |
Data-driven decision making and resource allocation |
ISO 45001 Clause 6, COR Element 2 |
|
Support (Resources, Competence) |
AI training simulations (VR/AR) |
PIXO VR, 360 Immersive, Synthesia video generator |
Competency assurance, immersive onboarding |
ISO 45001 Clause 7.2, COR Elements 4, 5 |
|
Operational Control |
Wearables & real-time monitoring |
Kenzen, Blackline, Modjoul |
Fatigue, heat stress, ergonomics tracking |
ISO 45001 Clause 8.1, COR Elements 6, 7 |
|
Emergency Preparedness & Response |
AI-generated emergency drill plans |
Tango AI, Custom GPTs |
Rapid SOP creation and drill evaluation |
ISO 45001 Clause 8.2, COR Element 8 |
|
Performance Evaluation (Auditing) |
AI-assisted auditing & inspections |
iAuditor, EcoOnline, Intelex. Q5, ChatGPT |
Streamlined inspections, repeat issue detection |
ISO 45001 Clause 9.2, COR Element 9 |
|
Incident Investigation |
NLP tools analyzing reports & trends |
ChatGPT, Power BI, custom bots |
Identifies root causes across incident data |
ISO 45001 Clause 10.2, COR Element 10 |
|
Continuous Improvement |
AI trend analysis and feedback loops |
Cority, Enablon, GPTs |
Drives prevention and change through insight |
ISO 45001 Clause 10.3, COR Element 10 |
|
Documentation & Records |
Auto-generated safety documents |
AirTable+GPT, Notion AI, Tango |
Ensures up-to-date and compliant documentation |
ISO 45001 Clause 7.5, COR Element 3 |