BackFive Stars for incident investigation that don't blame the worker.
The TruStar Method is a modern incident-investigation framework built specifically to address what experienced safety practitioners have long complained about in older root cause analysis tools. It runs on every TruStar RCA investigation and is supported by an independent AI bias audit on every analysis.
The shape of a complete investigation
Every causal factor in a TruStar investigation maps to one of five Stars. The Stars exist because most incidents are not the result of a single failure — they are the result of conditions, decisions, and signals across the whole work system that combined to make the incident possible.
Why a reasonable person would do what they did
The actions, decisions, and behaviours of individuals involved in the incident. The TruStar Method explicitly resists framing this as "human error." A reasonable, competent person did what made sense to them in the moment. The investigator's job is to understand why.
Older methodologies often stop here, finding "operator error" and recommending retraining. The TruStar Method treats this star as the beginning of the analysis, not the end. Every People finding triggers questions about the Conditions, Systems, Leadership, and Defenses that shaped what the person did.
The physical and environmental reality at the scene
Equipment, materials, environment, layout, weather, lighting, noise, ergonomics. The tangible things the person was working with and around.
A worn part, an obstructed sightline, a poorly designed control, or extreme weather can make a safe action mechanically difficult or a hazardous one easy. Conditions are often invisible to the worker performing the task daily.
The procedures, training, and processes meant to guide the work
The written and informal rules of the workplace. Procedures, training programs, JSAs, permits, communication protocols, and the gap between what the documents say and what people actually do day-to-day.
Most incidents involve a system that was either inadequate (the procedure didn't cover the scenario), inaccessible (the worker couldn't find the relevant rule), or unrealistic (the procedure was technically correct but operationally impossible to follow). The TruStar Method treats this gap as a first-class causal factor.
The organizational signals about what really matters
How the organization signals priority. Production pressure vs safety, resource allocation, supervisor visibility, the response to past concerns, and the messages workers receive about what gets rewarded and what gets ignored.
This is the star most commonly missing from traditional investigations. A crew that "rushed" did so because rushing was tacitly rewarded. A worker who "bypassed" the lockout did so because the alternative would have missed a deadline that mattered to someone above them. The TruStar Method requires the investigator to ask whether organizational signals contributed, and the impartial-assessment pass explicitly flags investigations that omit this star.
The barriers that were supposed to catch this and didn't
Every workplace has layers of protection: engineering controls, administrative checks, PPE, supervision, peer review. When an incident happens, by definition some of those defenses failed. This star asks which ones, and why.
Inspired by James Reason's Swiss cheese model. Most serious incidents involve multiple defenses failing in sequence rather than one catastrophic failure. By analyzing which defenses were missing, present-but-bypassed, or present-but-ineffective, the investigation reveals systemic vulnerabilities that wouldn't be visible from the immediate causes alone.
Six things the TruStar Method does that older approaches don't
Each of these came from a real, well-documented criticism of how incident investigation has historically been done in industrial safety. They are baked into how the TruStar Method works at every step.
The AI extracts every fact from your documents before any causal hypothesis is offered. The TruStar Method does not start by asking "which category does this fit?" — it starts by understanding what actually happened, then asks which of the five Stars the evidence implicates.
Compared with traditional approaches: Older tree-based methods often begin by asking the investigator to classify the incident before the evidence has been fully reviewed, which biases the analysis toward whichever branch is selected first.
Most incident investigations end at "the worker did X." The TruStar Method requires the investigator to consider what organizational signals, leadership decisions, and systemic conditions shaped that action. The Leadership star is mandatory in every causal analysis.
Compared with traditional approaches: Methodologies developed in the 1980s and 90s were heavily weighted toward individual human performance and equipment reliability, reflecting the engineering culture of that era. Critics have noted for years that this underweights organizational causes.
After the causal analysis is complete, a second AI pass reads the findings and explicitly checks for common reasoning traps: hindsight bias, individual-blame framing, premature closure, omission of organizational factors. The result is a paragraph the investigator must read and acknowledge before sign-off.
Compared with traditional approaches: Traditional methodologies have no equivalent. Once an investigator walks down a tree branch, there is no systematic check on whether the branch was the right one to walk down.
Every proposed corrective action carries its hierarchy-of-controls tier (Elimination, Substitution, Engineering, Administrative, PPE). The system flags CA sets that lean heavily on lower-tier controls and surfaces causal factors that no CA addresses. This is the AI doing peer review on the action plan before sign-off.
Compared with traditional approaches: Older methods often produce a list of corrective actions without weighing their structural effectiveness. The same action plan that lists "retrain the worker" alongside "install a guard" treats them as equivalent. They are not.
When a corrective action is marked verified, the system automatically schedules three follow-up reviews. The AI generates specific questions tailored to that CA, its causal factor, and its verification method. The reviewer answers whether the original problem has recurred and whether the control is still in place. This is what "closed loop" actually means.
Compared with traditional approaches: Most investigation processes measure success by whether CAs were completed on time, not by whether they actually prevented recurrence. Completion is not effectiveness.
The TruStar Method is designed to be productive on day one. The AI handles the methodology, the categorization, the bias check, and the action review — the investigator focuses on knowing their workplace and answering the AI's questions accurately. This makes serious-grade incident investigation accessible to organizations that don't have a dedicated safety department.
Compared with traditional approaches: Some popular incident investigation tools require a multi-day formal training course before users are effective with them, and ongoing reference to thick dictionaries during every investigation.
When the TruStar Method is the right fit, and when it isn't
- Near misses, first aids, recordable injuries, lost-time injuries
- Property-damage-only and environmental incidents
- Repeat-pattern investigations where prior CAs didn't hold
- Construction, oil and gas, manufacturing, mining, agriculture, healthcare, utilities, transportation
- Organizations that want defensible documentation without a 5-day training requirement
- Catastrophic process-safety incidents in heavily regulated industries (nuclear, refining process safety) where a formal multi-week, multi-investigator team approach is mandated
- Pure quality investigations where the issue is product defect rather than worker safety (other methodologies are more specialised here)
- Cybersecurity or IT-availability incidents (different domain entirely)
TruStar RCA also supports SCAT, ICAM, HFACS, 5 Whys, and Bowtie. If your organization has a mandated methodology, you can use it in this product.
The short version
The TruStar Method is what you'd build if you started incident investigation from scratch in 2026, with everything we now know about cognitive bias, organizational accident theory, hierarchy of controls, and how AI can act as a tireless peer reviewer.
It isn't a tree of pre-defined causes you walk through with a dictionary. It's five categories that cover the whole work system, an AI that proposes causal factors based on what the evidence actually shows, and a second AI that checks the first one's work for bias.
Try it on a real investigationBack to home