Salesforce AI Specialist Practice Exam

Training a lead conversion model with data skewed towards leads with low engagement is not recommended because such data does not represent the characteristics or behaviors that are most indicative of successful conversions. A model trained on this type of data may struggle to learn the relevant patterns and insights that drive effective lead conversion, leading to poor performance in actual lead scoring and conversion tasks.

Using historical data from fully converted leads offers a rich source of information about successful conversions and helps the model understand what factors contribute to lead success. Recent data from a prior sales quarter can provide context on current market conditions and lead behaviors that are more relevant to the model’s predictive capabilities. Combining data from both converted and non-converted leads can present a comprehensive view of the lead landscape, allowing the model to learn from both successes and failures.

In contrast, focusing on low engagement leads skews the training process towards examples that are less likely to convert, potentially resulting in a model that undervalues stronger leads or misidentifies high-potential contacts.