The business world has long understood the principle of “garbage in, garbage out.” Dan Federoff, VP and Head of Data at Bridgenext, says agentic AI has created a more urgent version: “garbage data, garbage agents”. Unlike a human analyst, an AI agent won’t pause to question bad numbers, it will act on them before anyone catches the mistake.
In this TechTarget feature, Federoff explains why strong data governance, standardized prompting, and clear engineering guardrails are essential for scaling AI agents safely without introducing new risk.