A machine-learning-based data security service would best meet the pharmaceutical company's requirements to discover data, monitor changes, and identify suspicious activity across thousands of independent resources in cloud object storage. Machine learning can analyze vast amounts of data, detect patterns, and alert administrators to anomalies or suspicious activities without manual intervention. Traditional file integrity monitoring or data classification might not scale well or adapt dynamically to the complexity and size of the company's environment.