A sales operations lead at a mid-size manufacturer wants to enable predictive lead scoring in Dynamics 365 Sales, but the organization has only about 5,000 closed leads with recorded outcomes from the last six months, well below the volume needed to train a reliable org-specific model. What should the consultant expect to happen when the feature is turned on?
Select an answer to reveal the explanation.
Short Explanation
Think of this like a new employee who has not yet seen enough deals at this specific company to develop instincts, so they borrow the collective experience of peers at similar companies until they build up their own track record. That is exactly what happens here: with only a few thousand outcomes on file, there is not enough of this company's own history to teach a model its unique patterns, so the system leans on a broader, anonymized pool of similar-organization data as a starting point. It keeps working, just with borrowed judgment rather than home-grown judgment, and it shifts toward the company's own patterns as more real outcomes get recorded. The traps here are assuming the feature simply stops working below some threshold, assuming a human has to hand-code scoring weights instead, or assuming the system will go fetch more leads from another app to pad the numbers. None of those match how the model actually adapts to limited data.
Full Explanation
The correct answer is A. Predictive lead scoring does not require an organization to have massive historical volume before it can be used at all; when an organization's own qualified and disqualified lead history is too thin to train a dependable custom model, the feature uses a generic model built from pooled, anonymized patterns across comparable organizations, and it transitions toward an org-specific model as more of the client's own outcome data accumulates. Option B is incorrect because the feature is designed to remain usable below the ideal data threshold rather than blocking activation entirely. Option C is incorrect because predictive lead scoring is a machine-learning capability that learns from outcome patterns; it is not a manual rule-weighting engine, which is a separate, rules-based approach used elsewhere in the platform. Option D is incorrect because the platform does not automatically pull additional lead records from Marketing to artificially inflate the training set; data volume grows only through the organization's actual ongoing lead activity, and cross-application record importing is not part of how the scoring model sources its training data.