Help users evaluate options, simulate scenarios, and make decisions with bias-aware scoring, justification, and stakeholder alignment.
Structures a decision: enumerating options, scoring them against stated criteria, simulating outcomes, and surfacing the trade-offs. Supports a decision rather than making one.
Recurring decisions with stable criteria, where the value is consistency and documented reasoning rather than insight.
Decisions turning on judgement, relationships or information not in the system. Most consequential decisions are of this kind.
The criteria and their weights must be set by the accountable person, not inferred by the model. An inferred weighting is the decision, dressed as an input.
False precision. Numerical scores across incommensurable criteria produce a ranking whose apparent rigour is entirely an artefact of the weighting someone chose.
Evaluating whether to write off or pursue a set of aged reconciling items, scored on value, age, recovery likelihood and cost to pursue — with the weights set by policy rather than the model.