A consultant is implementing predictive lead scoring in Dynamics 365 Sales for a marketing team that wants to prioritize which inbound leads should be handed to sales reps before those leads become opportunities. Sales leadership asks whether the same predictive opportunity scoring model used later in the pipeline can simply be reused for this purpose. What should the consultant explain?

Predictive lead scoring and predictive opportunity scoring are separate models trained on different data, engagement and firmographic signals versus in-flight opportunity activity, so a distinct lead-scoring model is required. Predictive lead scoring evaluates leads using engagement signals such as email opens, website activity, and firmographic attributes, since a lead has no deal-specific activity yet. Predictive opportunity scoring evaluates opportunities using pipeline-stage activity, deal characteristics, and historical win/loss patterns from closed opportunities. Because the inputs and the outcome being predicted differ, the two use distinct models rather than one being repointed at the other entity. The claim that the opportunity scoring model can be pointed at the leads entity by changing a single configuration setting is incorrect because there is no configuration switch that redirects the opportunity model onto lead records; the entities carry fundamentally different data. The claim that predictive scoring only applies once a record becomes an opportunity is incorrect because Dynamics 365 Sales does offer predictive lead scoring as a distinct capability, so leads are not left unscored before conversion. The claim that lead scoring requires a separate Copilot license from opportunity scoring but once licensed uses the identical underlying model is incorrect because the two features are not the same model wearing different licensing; they are separately trained models addressing different questions in the funnel, and licensing terminology does not change that underlying distinction.