LEODSI — a framework for learning before the next incident forces you to

You could ask if most organisations learn from incidents. The investigation is commissioned, the report is produced, the root cause is identified, and the corrective actions are assigned. Then, for the most part, operations continue. What did they learn? What have they have changed?

We need to look deeper if we want to move beyond single-loop learning. We need to ask deeper questions. What generated the incident — the interactions, pressures, adaptations, and systemic conditions that shaped the decisions that were made — is rarely understood in full, and rarely addressed at the level where it would actually produce change.

LEODSI — Learning from Emergent Outcomes in Diving Systems and Interactions — is a framework developed by The Human Diver that takes a fundamentally different approach. It begins with a different question. Not 'what went wrong and who is responsible?' but 'how do we design the system so that good performance is easier, poor outcomes are less likely, and learning is normal?' That shift — from fault-finding to system understanding — changes everything that follows.

Although LEODSI was developed in and for the diving world, its intellectual foundations are drawn from human factors, systems thinking, and the Safety-II tradition that has application across every high-risk industry. The framework is built for organisations that want to understand performance before an adverse event forces the question — which makes it directly relevant to any leader trying to improve capability proactively rather than reactively.

LEODSI is not limited to incidents. It is equally applicable when things go right, when outcomes are benign but uncomfortable, when adaptations prevent escalation, and when performance varies under pressure. Learning does not require a catastrophe as its trigger.

The system, not the individual

The central analytical tool within LEODSI is PETTEOT — Person, Environments, Tasks, Tools and Technology, External Influences, Organisation, and Time. These seven elements make up the work system in which any task is performed. People sit at the centre of that system not because they are the problem, but because they are the point where all the other elements come together. What any individual does at any moment is shaped by all of the elements surrounding them, and understanding performance — good or poor — requires examining all of them.

This matters because the standard approach to investigation does the opposite. It identifies what the individual did, applies hindsight to judge whether that decision was reasonable, and assigns responsibility accordingly. LEODSI's PETTEOT framework asks instead: given the environment, the task demands, the tools available, the external pressures, the organisational conditions, and the time constraints that person was operating under — why did their action make sense to them in the moment? That question is not about excusing poor performance. It is about understanding it well enough to change the conditions that produced it.

The seven elements in practice

Person includes the skills, experience, cognitive and physical state, mental models, and team dynamics of everyone involved — not just the individual closest to the outcome. Environments covers the physical and sensory conditions in which the work is done. Tasks addresses the activities required, their sequencing, complexity, and how they stack and compete for attention. Tools and Technology encompasses all equipment, systems, and artefacts used — including how they are configured, maintained, and understood. External Influences captures the factors outside the team's immediate control that shape how work is planned and conducted: weather windows, regulatory requirements, budget constraints, travel logistics. Organisation addresses the structures, norms, cultures, and informal practices that shape how work is organised.

Time is the seventh element, and it deserves particular attention because it is the one most consistently overlooked. LEODSI distinguishes between three temporal perspectives:

• Work-as-Imagined — the timeline that should happen according to the plan;

• Work-as-Normal — what actually happens when things go routinely; and

• Work-as-Done-on-the-Day — what actually happened on this specific occasion, and why it differed.

The gap between these three timelines is often where the most important system intelligence lives. Compressed briefings, skipped debriefs, rushed preparations — these are temporal decisions, and they are almost always driven by systemic pressures rather than individual carelessness.

Learning from what goes right

One of LEODSI's most important characteristics is its explicit application to successful outcomes, not just adverse ones. The vast majority of operations in any high-risk environment conclude without incident — not because the system is perfectly designed, but because people are continuously adapting, compensating, and absorbing variability. That adaptive performance is the real intelligence of the system. If your learning processes only activate when something goes wrong, you are missing most of the information that would help you understand how your system actually works and where its genuine vulnerabilities lie.

LEODSI's After Action Review (AAR) structure — built around four questions — is designed to be used routinely, not just after incidents. What were we trying to achieve, and what actually happened? What went well, and why? What could have gone better, and why? What will we do differently? Those questions, asked consistently after routine operations as well as difficult ones, build an organisational intelligence that no incident investigation programme alone can provide.

What this requires of leaders

LEODSI is explicit that the framework only functions in an environment of psychological safety and Just Culture. Learning does not happen when honesty carries professional risk. The ethics of the learning environment are not an add-on — they are, as the LEODSI documentation puts it, the infrastructure that makes learning possible. Leaders who want to use this framework need to create the conditions in which people can describe what actually happened — including the adaptations, the shortcuts, the moments of uncertainty — without fear that honesty will be used against them.

The organisations that get this right are those that treat every outcome — successful, uncomfortable, or adverse — as information about how their system is functioning. They build learning into the rhythm of operations rather than treating it as a response to crisis. They understand that the question 'why did this make sense at the time?' is the beginning of genuine improvement, not an evasion of accountability.

LEODSI exists to help organisations answer that question systematically, continuously, and before circumstances force them to.