While agentic AI is a very promising tech, we also have to understand that agents have a ton of limitations including the fact that they often do not learn from their mistakes and hence human involvement is important to prevent errors from compounding. https://t.co/G5aIK8vgEA