Risk assessment is a judgement call. The problem is that judgement is inconsistent — the same risk, assessed twice, rarely produces the same answer. That gap is where liability lives.
Rysc Ai structures the assessment so the scoring is repeatable, the reasoning is documented, and the result is something a practitioner can explain without referring to a manual.
Traffic-light registers collapse professional judgement into a colour. LES scoring with fuzzy number bounds structures that judgement — shared language, documented reasoning, a result that holds up to scrutiny.
AI-generated narrative from live project data — threshold watch, treatment financials, workflow status — formatted for client submission. From a complete register to a deliverable in one click.
Specialist modules triage by project type and location. The platform applies what's relevant and flags what might be missing — before it becomes a gap in your submission.
ISO 31000 requires risks to be identified systematically. Cause → consequence structure ensures consistent descriptions that can be compared and aggregated across the register.
The standard requires analysis using 'consistent criteria'. TFN decomposition into three components with defined scales produces consistent, comparable results across assessors and projects.
Evaluation means comparing risk against defined criteria to decide what needs treatment. Three zones with numeric boundaries (0–60, 60–110, 110+) give each risk a clear action category.
Treatment selection requires weighing cost against effectiveness. Quantitative cost modelling supports evidence-based treatment selection at committee level.
Version history and aging analysis make stale risks visible so they get re-assessed and closed. The state machine enforces review at each workflow transition.
The standard requires recording and reporting as part of the risk management process. Preformed dashboards and charts provide the visual evidence that your process is working.
The same risks recur across projects. A searchable lessons library means the next project team starts with institutional knowledge, building on what was learned before.
Anchoring, recency bias, and unfamiliarity effects are well-documented in risk assessment. Structured decomposition and range-based inputs are practical countermeasures built into the scoring workflow.
Most platforms score a risk once, in one domain. But a single event — a contractor default, a structural failure, a reputational incident — carries simultaneous exposure across safety, cost, time, quality, and reputation.
Rysc Ai calculates a compound vector score — the Euclidean magnitude across all active domains. Adjust the sliders to see how domain scores combine into a true composite exposure.
Every possible combination of Likelihood, Exposure, and Severity — plotted in three dimensions. Drag to rotate. Hover any point to inspect its score. The envelope makes the scoring model transparent.
Likelihood, Exposure, Severity scored in shared language across six descriptive levels. S^1.2 severity weighting. Triangular fuzzy number bounds capture the range of professional judgement — optimistic, most-likely, pessimistic. Three-zone thresholds drive workflow automatically.
RyscAid does more than fill in values — it challenges the assessment. Describe a cause and consequence, and RyscAid identifies exposure domains the assessor may not have considered, asks whether adjacent risks have been captured, and recommends LES values with plain-language reasoning. It acts as a second opinion before the entry is committed — reducing the gaps that come from working in isolation.
AI-generated narrative with threshold watch, top risks, treatment financials, and workflow status. The report speaks in outcomes — what is at risk, why it matters, what has been done. No LES notation exposed in client output.
Schema-per-tenant database isolation. Role-based access — PM, Moderator, Risk Manager. State machine workflow with PM approval. Compound vector scoring across domains. Immutable audit trail. HSWA s23/s24 notifiable risk triage on every entry.