Most decisions go wrong before anyone weighs the options — in how the decision gets made and who gets to make it. These are small, transparent helpers for that meta-decision. Pick the one that fits the choice in front of you.
Answers: which decision-making method fits this decision? — from Autocratic through Advice and Consent to Consensus — and which decision-rights tools (RACI, RAPID, DACI, DARE) fit it or quietly break it.
Tune a decision on seven dimensions — urgency, reversibility, predictability, interdependence, where the knowledge sits, agreement, and buy-in — and get a best-fitting method with reasons. The highest-order helper: use it first.
Open →Answers: should this be a workflow, AI within a workflow, or an agent — and how much freedom should that agent have?
The Agency Design Canvas. Describe the work, the value of independent action, the consequences of mistakes and the controls you have in place; get a suggested architecture, a permissible-autonomy ceiling and human-oversight guidance. A decision aid, not permission to deploy.
Open →Other domain-specific decision helpers can slot in here as they are built — each one a single, transparent page in the same family.
What these are. Each helper is a self-contained, client-side page — no accounts, no tracking, and nothing you enter leaves your browser. They are structured reasoning aids built on published research and standards, not validated instruments or authorisation to act. Treat every recommendation as a prompt for judgement.