Insights from Our Exclusive Survey

It’s one thing to sit in a boardroom and talk about the "disruptive potential" of AI in Health & Safety, but it’s another thing entirely to get out on the floor and see how this new technology is actually landing in the hands of the modern workforce.

Over the last couple of years, the landscape surrounding AI has undergone a seismic shift, moving so fast it’s becoming increasingly difficult to tell where the genuine breakthroughs end and the corporate illusions begin. 

If you scroll through LinkedIn or catch the evening news, the media would have you believe that every person and their dog are currently implementing complex AI systems into their daily workflows. We’re being sold a vision of a world where AI is already the co-pilot for every safety decision made.

But how much of this is reality, and how much is just headlines?

Our Survey

To find the real pulse of the industry, we recently conducted a deep-dive survey of 223 safety leaders and executives across New Zealand, Australia, and the UK. We didn’t just want the "C-suite" view; we wanted the perspective of the people closest to the day-to-day decisions. Over 70% of our respondents were Health & Safety Managers or Leaders.

These respondents represented the full spectrum of industry, from lean, small teams to massive enterprises, with 30% coming from companies with more than 1,000 employees in high-impact sectors like manufacturing, transport, energy, and government.

The results were a wake-up call. The data suggests that while everyone is talking about AI, very few have actually moved it into the core of their operations.

The Great "Exploring" Phase

Currently, 58% of organisations report that they are still firmly in the "exploring" phase. In plain English, "exploring" usually means people are playing around with tools like ChatGPT in a fragmented, ad-hoc way. It’s the digital equivalent of kicking the tires. This experimentation is happening outside of formal company systems, often on personal devices. 

While curiosity is great, this unstructured approach is a safety risk in itself. When AI usage is fragmented, you end up with a trifecta of organisational headaches:

No Governance: No one knows what data is being fed into these models or where it’s going.

No System Integration: The AI isn't "talking" to the rest of your system; it’s an island.

No Consistency: You have five different supervisors getting five different answers because there is no standardised prompt or process.

Beyond the explorers, 20% of organisations have started piloting specific features — moving from "what is this?" to "let's see if this specific tool works." However, there’s a notable 10% who are heading in the opposite direction, working in restricted environments where AI is currently banned or so tightly controlled that it’s effectively unusable. 

The Implementation Gap

The most telling statistic from our research is that the jump from "playing with tools" to an "integrated daily workflow" is where most organisations are currently stuck.

Only 4.5% of respondents have managed to truly embed AI into their day-to-day safety processes. Think about that: despite the billion-dollar valuations and the constant media cycle, fewer than 5 in 100 safety teams are actually using it in a structured and consistent way.

Even in large enterprises (1,000+ staff), where you might expect deeper pockets and more specialised digital transformation teams to drive movement, that figure only hits 10%.

So why the stagnation? 

Because success isn’t picking a shiny new vendor or buying a few dozen licenses, it’s doing the "boring" but essential hard work of configuration and change management. 

If the technology doesn't match your specific, granular safety goals, it becomes a distraction. If the frontline doesn't feel engaged the project risks becoming just another expensive statistic in the "failed implementation" column.

AI Needs a Strong Foundation

A significant part of this challenge is ensuring H&S leaders are using AI on an unstable foundation when it comes to systems. It’s common to find a spreadsheet for hazards, a standalone app for incidents, and an entirely separate database for contractor management.

So when datasets aren’t talking to each other, the AI is forced to make guesses about how they relate, and those assumptions are often wrong. The most valuable safety insights don't live in a single report; they live at the intersection of these different data points. 

Without a unified system, we risk building a high-tech skyscraper on a burning foundation.

The Tide is Coming In

Despite the current caution, we aren't going to stay in this holding pattern for long. The survey results show a massive psychological shift on the horizon. Three-quarters of leaders (74.9%) expect AI adoption to increase over the next year, with almost one in five bracing for a "significant increase."

The "wait and see" period is ending. The industry is realising that while the early hype was inflated, the underlying utility of the technology is real. We’re standing on the edge of a major step-change. 

The era of cautious, fragmented experimentation is dying, and the era of integrated, high-value safety tech is beginning. 

Leveling Up the Right Way

As we move into this next phase, the challenge for H&S leaders is “how do we get AI right?"

To avoid the pitfalls of the 58% who are stuck in the exploration phase, organisations need to move toward structured implementation. This means:

Defining the Use Case: Stop asking "What can AI do?" and start asking "What is our biggest safety bottleneck that a machine can help solve?" (e.g., analysing 5,000 near-miss reports for hidden trends).

Governance First: Establish clear guidelines on data privacy and ethical usage before the tools are rolled out.

Human-Centric Design: AI should remove the "drudge work" of safety to free up safety professionals to do what they do best: engage with people on-site. 

Final Words

We’re moving toward a future where AI won't just be a chatbot you talk to when you're bored; it will be the engine that helps us manage risks before they become incidents and ensures that no critical piece of safety data ever falls through the cracks.

The goal has never changed: we want to get people home unharmed. The tools are just getting smarter. The challenge now is ensuring that when you do decide to "level up" your tech stack, you’re building on a foundation that actually works for your people, rather than just adding another layer of digital noise.

Key Takeaways

• The "Exploration Phase" is widespread: 58% of organisations are currently experimenting with AI in an ad-hoc, unstructured way. Without formal governance or system integration, this fragmented approach creates significant risks, including data privacy concerns and inconsistent safety advice across the workforce.

• The Implementation Gap: Despite the hype, only 4.5% of safety teams have fully embedded AI into their daily processes. Even in large enterprises, successful integration remains low because the "boring" work of change management and technical configuration is often overlooked in favor of chasing shiny new tools.

• The "Wait and See" era is ending: A psychological shift is imminent, with 75% of safety leaders expecting adoption to increase significantly over the next year. To "level up" successfully, leaders must shift from asking what AI can do to identifying specific safety bottlenecks that AI is uniquely qualified to solve.

To further your organisations health and safety management, check out our Health & Safety Management Software and book a demo today!