Published 12 days ago • loading... • Updated 11 days ago
45% of execs limit human AI oversight to high-stakes work, or don't have any oversight at all
The survey found many leaders are reducing human review as companies deploy AI, and nearly one-third of executives say they trust AI to make most decisions.
A Zapier survey of executives, conducted between July 1 and July 7, 2026, reveals that many respondents provide AI with full, unsupervised autonomy for various tasks.
Unsupervised AI has already cost organizations time, money, or reputation, according to the report. These setbacks reveal the critical risks of deploying automated systems without established governance policies.
While companies prioritize speed, many lack sufficient oversight. Executives allow AI to handle high-stakes decisions, including budget approvals and vendor management, despite the potential for costly operational failures.
Repeated errors often stem from "rubber-stamping" AI outputs without human oversight. Some executives permit AI to update customer records automatically, causing significant compliance and operational issues.
Experts recommend implementing governance frameworks that require human approval for financial actions to prevent future losses. Organizations must balance AI speed with strategic oversight to ensure automation supports critical judgment.