Agent-washing

Agent-washing — a term coined by Gartner in 2025 — is the practice of rebranding existing software, chatbots, and plain automation as “agentic AI” because agents are what buyers want to hear. Gartner’s estimate at the time: of the thousands of vendors making agentic claims, only around 130 offered genuine agent capabilities.

Why it matters: Buyers try to evaluate the claim on vibes when there’s a four-part test: Does it run autonomously without per-step prompting? Does it coordinate genuinely separate agents on a shared task? Does it produce output you can verify without redoing the work? Does it persist state across sessions? Fewer than three yeses is a chatbot in a costume — and asking a vendor to demonstrate the four properties is informative precisely because many can’t. The label costs nothing; the properties can’t be faked in a live demo on your data.