An administrator is preparing to turn on predictive opportunity scoring for the first time in a new Dynamics 365 Sales Premium environment. (Select TWO.) Which two conditions must be true for the model to begin producing usable scores?
Select all correct answers, then click Submit.
Short Explanation
Getting opportunity scoring working really comes down to two things happening together. First, somebody has to actually turn the feature on and point it at the opportunity data, since it is not running in the background by default. Second, the model needs enough closed deals in its history, a real mix of wins and losses, to have something to learn from, the same way any prediction only gets good once it has seen enough real outcomes to compare against. Flip either one off and nothing useful comes out the other end. What trips people up is assuming some unrelated field, like which territory a deal belongs to, is secretly required, when really the model does not care about that at all. Same with the idea that you have to shut off classic forecasting first, as if the two features would collide, when in reality they run independently and there is no reason to disable one to use the other. Configuration plus history is the whole story here.
Full Explanation
The correct answers are A and B. Predictive opportunity scoring is a machine-learning capability that must be explicitly enabled and pointed at the Opportunity table through Sales Insights settings, and like any predictive model it also needs a meaningful history of closed opportunities, spanning both won and lost outcomes, to learn what separates a likely win from a likely loss. Without the configuration step nothing runs at all, and without enough labeled history the model has nothing to learn patterns from, so both conditions are required together rather than either alone being sufficient. Option C is incorrect because sales territory is not a prerequisite field the scoring engine depends on; opportunities without a territory assigned can still be scored so long as the model has the signals it actually relies on. Option D is incorrect because predictive opportunity scoring and classic forecasting are separate capabilities that can coexist; there is no requirement to turn off forecasting before enabling AI-based scoring, and doing so would remove a tool leadership may still want to use. The genuine prerequisites are the explicit feature configuration and a sufficient base of historical outcomes, not unrelated data fields or disabling other features.