Why We Fail to Learn—and How to Fix It
Learning from incidents
Incidents are costly and harmful yet also powerful opportunities to improve and ensure a similar event with the same root cause doesn’t recur — however we often miss these opportunities. It’s not because of a lack of care that organizations fail to learn, rather it is shortage of time, our own biases, or cultural norms that can lead to shallow insights and quick fixes. When we don’t learn, our barriers remain weak, and history repeats. A strong investigation paired with an energetic process of reflection does more to prevent recurrence than any noticeboard or pep talk.
Why investigations fall short
Under pressure to “close the case”, teams rush actions or even drag out inquiries until the momentum fades. Sometimes the basics are missing — injured parties are unavailable, sites restricted, or evidence thin — so we settle for what’s visible or familiar or jump to “wish-list” actions unrelated to real causes.
A tiered approach
Many organizations invest a huge effort in exchange for limited learning. A “one size fits all” process — team, meetings, standard method, actions, upload, done. This can waste time or produce a “tick-box” culture where the wrap-up (not the learning) becomes the goal. Separating incidents into risk-based tiers can direct the effort to where it matters most: go deep when risks and learning potential are high; use light methods for minor issues. Whatever the approach, keep it simple (KIS), as far as possible.
The right tools – often simpler the better
Unnecessarily complex methods get in the way of quickly identifying the underlying reasons incidents occur. They can also be overwhelming for practitioners not regularly working with incidents. We can work with various techniques but, given a blank slate, find root cause analysis (RCA) for serious events and 5 Whys for simpler ones strikes a good balance.
Sakichi Toyoda created the 5 Whys after seeing workers fix symptoms, not causes, of problems — his work offers us a quick way for anyone to uncover underlying process failures and drive lasting improvement. In essence, convening a small group of people with first-hand experience and different systemic insights to ask “why did this happen?” and asking that not once but five times, allows
For example, a chemical leak occurs – why? The pipe broke – why? Corrosion – why? a corrosive gas composition in the pipe – why? Has a new reagent been introduced? Why? To improve product safety. Rapidly equipped with the above information, we might promptly move onto consider questions of management of change, inspection or sampling.
The approach is far from perfect but is without doubt fast and simple – factors that help our frontline workers stay engaged and on top of our incident workload.
Curiosity can serve us better than expertise
Persistence and curiosity are traits vital in investigation work. By asking “what would have to be true for this hypothesis to hold” we move from a belief based investigation to an evidence-based investigation, from assumptions based on expertise to conclusions based on fact.
But that’s not to say that investigation does not need practice – the tools we gain from training and experience become blunter without practice. If full-time investigators aren’t feasible, develop a small, active cadre and coordinate their involvement in incidents — this will be far more effective than a large pool of trainees that lie idle.
Share the learnings: Don’t post it, host it
Too often, when communicating “lessons learned” these are emailed or posted, not lived. Think of learning as a hosting occasion, not a broadcast. Own the gravity of what has happened, or – in the case of a near miss – nearly happened and the conviction to prevent recurrence. Who is your audience, what do you want them to know, feel, and do differently? With limited time, focus on fewer lessons and explore them deeply.
- Go deep, go narrow: communicate lessons on fewer incidents, but do them well.
- Aim for the lesson, not the headline: high-profile cases aren’t always the most insightful.
- Target your audiences: tailor messages and materials.
- Create a toolbox: a mix of repeatable, easily identifiable tools for alerts, campaigns, and reflection.
- Make lessons actionable: create visual guides and reflective questions that can be used for toolbox talks or walkarounds.
- Create energy: Consider making competitions, visuals or props that help catch attention
- Communicate: Reinforce real-world engagements with simple, powerful visuals and messages in various media
How AI Can Transform Learning Transfer to the Frontline
Today, AI gives us a major opportunity to bridge the gap between “knowing” and “doing.”
At SnSD, we are already applying this through our in-house built AI engine, trained on our vast internal knowledge base, investigation learnings, safety cases, Risk insights, and decades of real industry experience. Instead of long reports or dense PowerPoint decks, AI helps convert critical lessons into nugget-size, frontline-ready communication materials — simple visuals, toolbox prompts, short reflective questions, and 60-second summaries that supervisors can use on the spot.
This dramatically increases the likelihood that insights from investigations are not only understood but applied in day-to-day work. As AI matures, it will allow organizations to personalize lessons for different roles, languages, and contexts — ensuring that learning finally reaches the people who need it most.
The takeaway
Effective learning from incidents comes from combining pace with curiosity: dive deep where you must and move quickly where you can, to understand the conditions that allowed the event. Apply process and rigor rather than jumping to assumptions or familiar fixes. Establish proficient investigators and a simple mix of tools to stay on top of incident workload and get to genuine system weaknesses. Take the time to communicate wisely, not just widely and deploy AI to multiply the impact at the frontline.
