An organization is choosing between a pre-trained API and a custom ML model for a common document-processing task. Which consideration most strongly favors trying the pre-trained service first?

1. A guarantee that every specialized business rule is supported
2. Complete elimination of data-privacy responsibilities
3. A guarantee that no evaluation is necessary
4. Faster delivery with less training-data and model-operations work