Salesforce recommends implementing guardrails and human review for AI agents, stating that agents should not make final decisions or disperse funds.
November 21, 2025
high
policy
Stated best-practice approach to deploying AI agents emphasizing human oversight and operational limits.