The answers by question key. Empty on failure.
OptionalErrorWhy the decision failed, when Success is false.
OptionalModelThe MJ decision model that answered, after any failover: Jev, LLM Decision
(AIDecisionRunResult.modelInfo.modelName, RunDecisionResult.ModelName).
OptionalPromptThe MJ: AI Prompt Runs row the decision wrote, when it wrote one.
OptionalResolvedThe exact model behind it, as the driver reports it: the vendor's dated model for a vendor
decision model, the chat model for LLM Decision (AIDecisionRunResult.DecisionResult.ResolvedModel,
RunDecisionResult.ResolvedModel).
Whether the decision ran and its answers were read.
A routing decision's result, as InterpretRoutingAnswers reads it.
RunDecisionResultin@memberjunction/graphql-dataproviderhas this shape.Its
ModelNameandResolvedModelsay which model answered. Calibration is per exact model (ROUTING_CONTINUES_CALIBRATION), so routing acts on the thread Likelihood only when both are reported and name a calibrated model.