An AI application development team has been given access to user information and now must...

In the AI data lifecycle, converting raw user information into model-readable form (e.g., cleaning, parsing, feature engineering, encoding, normalization/standardization, and schema alignment) is performed during data preparation. This phase establishes input quality and structure prior to training and inference and may include normalization as one of several preparation steps. Data collection acquires data, minimization limits data to what’s necessary, and normalization is a specific technique—not the overarching phase.