The number: in agentic AI, the first barrier to scaling isn't technology, it's trust

With generative AI, the risk was that the model said something wrong, and a human could catch it before acting. With agents, the risk is that it does the wrong thing: the action has already happened. That is why nearly two-thirds of organizations name security and risk as the top barrier to scaling agentic AI — ahead of regulatory uncertainty or technical limits — while only one in three has governance mature enough to keep up, and 80% have already seen an agent misbehave. For a small or mid-sized company, this shifts the first project: before automating a task the AI will carry out on its own, write down who checks what, at what threshold, and with what power to stop it.

Source: McKinsey, "State of AI trust in 2026: Shifting to the agentic era," 2026 (AI trust maturity survey, ~500 organizations, fielded Dec 2025–Jan 2026).