Home
Learning From Near Misses – Before They Happen: A word from AGuyIKnow

Learning From Near Misses – Before They Happen: A word from AGuyIKnow

Technology doesn’t fail in mining – adoption does.
We often use that line in the context of digital transformation, but the same principle applies to safety. In high-risk environments, people don’t rely on procedures in the moment – they rely on experience.

And that’s where a gap exists.

The Limits of Traditional Safety Training

Most safety training is built around:

  • procedures and compliance
  • historical incidents
  • static scenarios
  • point in time, in person training

It’s necessary – but it has limitations.
Because when someone is standing near moving equipment or entering a blind spot, they’re not recalling a slide deck. They’re making a real-time decision under pressure.
That decision is shaped by what they recognise and what feels familiar.

The Opportunity in Near Misses

The Opportunity in Near Misses
Near misses are some of the most valuable learning opportunities in mining.
Widely recognised safety models, such as Heinrich’s Triangle, suggest that for every serious incident, there are many more near misses. This represents a significant opportunity to intervene before harm occurs.
Currently, learning from these near misses remains inconsistent.

You only learn from:

  • what has already happened
  • what has been reported
  • or what you’ve personally experienced

That leaves a gap – particularly for new workers, new sites, or changing operating conditions.

The Challenge With Delivery

Part of the challenge is how these learnings are delivered.
In practice, organisations are often forced to choose between two approaches.
Traditional methods rely on workshops and static content which is time-consuming and often disconnected from when learning is needed, often requiring workers to step away from operations. At the other extreme, immersive training requires full-scale video production or investment in VR hardware, both of which are costly, difficult to scale and can quickly become outdated.
The gap is clear: there has been no practical way to deliver realistic, experience-based learning at scale.

Making Near Misses Experiential

One of the most practical applications of Generative AI we’re seeing is the ability to recreate realistic operational scenarios.
Working with mining organisations, we have developed Near Incident Miss Scenarios (NIMS) which are immersive, scenario-based experiences built from real incident and near miss data.
These scenarios recreate the sequence of events, the decisions made, and the consequences that followed. Importantly, they use actual photos from the physical locations.
For example, scenarios can recreate events such as light-vehicle / haul-truck interactions, unplanned movement of heavy mobile equipment and live-conveyor proximity incidents. This allows workers to see how situations unfold in familiar environments and where the critical decision points are.
They are not abstract examples. They reflect real operating conditions and the types of situations people encounter on site.

From Instruction to Experience

Instead of telling people what to do, we can show them what it looks like when things go wrong in their physical environment and, importantly, where the critical decision points are that lead to negative consequences.

This helps build:

  • stronger hazard awareness
  • better decision-making
  • greater confidence in the field

Because people are not just being trained – they are being prepared for real risk scenarios they will encounter on the job. Their ability to respond is critical to ensuring the safety of workers on site.

Embedding Learning Into the Flow of Work

A key shift is not just how learning is created, but how it is accessed.
Traditionally, these learnings are delivered through workshops or training sessions that are time-intensive to coordinate and often disconnected from the moment they’re needed.
This creates a reliance on point-in-time learning, where retention can fade and access to information is limited once the session ends.
Rather than relying on one-off training sessions or documents stored in repositories, these scenarios can be delivered through existing tools such as Teams and mobile devices, and surfaced at the point of need.
That means they can be used in pre-starts, toolbox talks, onboarding, or accessed on demand in the field.
The result is learning that is not separate from work – but part of it.

Why This Matters Now

Mining remains a high-risk industry, with serious incidents carrying both human, financial and operational consequences. In Western Australia alone, Mining industry compensation claims exceed $200 million annually. Source: WorkCover WA
At the same time, operations are becoming more complex. New technologies introduce new behaviours, interfaces and failure points.
In many cases, incidents are not caused by a lack of procedures, but by how those procedures are applied in real situations.
If capability doesn’t evolve at the same pace, risk increases – regardless of how well the technology is designed.

Looking Ahead

This approach doesn’t replace traditional safety training – it strengthens it.
The opportunity is to move beyond documenting incidents and toward preparing people for them. Because in safety-critical environments, experience matters.
And the closer we can get to creating that experience before it happens, the better prepared people will be when it does. For those interested in seeing how this works in practice, AGuyIKnow will be showcasing Near Incident Miss Scenarios (NIMS) at Austmine GRX26 in the Innovation Zone.

About the Author

This perspective reflects work being undertaken by AGuyIKnow, a Perth-based consultancy specialising in capability, learning and digital adoption within mining and METS environments, focused on embedding safety, decision-making and operational capability into real-world practice. AGuyIKnow have recently formed an Education Partnership to support and promote the industry to young people, fostering the next generation of industry leaders.