A risk register is the living list of what could derail a project, scored so the team can focus on the threats that matter most. The hardest part is not the scoring — it is remembering the whole space of risks, including the unglamorous people and dependency risks that sink projects more often than technical ones. This builder turns your project details into a prompt that makes an LLM enumerate risks across every category, score them consistently, and draft mitigations for the worst.
How it works
You describe the project type, team size, delivery timeline, and any hard constraints (budget caps, fixed dates, regulatory requirements). You choose a scoring approach — the default is likelihood × impact on a 1-to-5 scale, giving a 1-to-25 risk score. The builder assembles a prompt that instructs the model to identify risks across technical, schedule, resource, scope, external/dependency, and people dimensions, score each one, rank by score, and then — for the top risks — propose a mitigation action, an owner role, and an early-warning indicator. The prompt is generated in your browser and nothing is uploaded.
Tips and examples
Be specific about constraints: “fixed launch date for a regulator demo, two of five engineers part-time” surfaces far more relevant risks than “software project.” Run the prompt early, when mitigations are still cheap — a risk register written the week before launch is a post-mortem in disguise. Always edit the AI’s scores with your team; the model cannot know that one vendor is historically unreliable or that a key engineer is about to take leave. Keep the register live: re-run the prompt at each milestone, retire risks that have passed, and add new ones as scope shifts, so it stays a decision tool rather than a one-off document.