06·"Agent washing" — How to Spot a Fake Agent

Introduction

The term "agent washing" describes the phenomenon where software vendors (or internal teams) attach the "AI agent" label to products that are in reality advanced chatbots, pipelines with hard-coded logic, or simple scripts with an LLM call. This lesson provides a set of diagnostic questions and criteria that distinguish a genuine agent (which plans autonomously, remembers, and acts) from a system that merely looks like one. This skill is essential for a developer who must evaluate a library, a SaaS offering, or a PoC before investing time in integration.