A survey by VentureBeat, published on October 8, 2026, reveals a decline in corporate confidence regarding autonomous AI agents. The percentage of companies permitting AI to implement production changes without human oversight dropped from 75% in July to 56% in August.
The August data is based on responses from 140 participants. The composition of the survey group shifted, with decision-makers accounting for 53% of respondents in August, compared to 44% in July. Among those responsible for final AI purchasing decisions, the willingness to allow automated changes fell from 88% to 61%.
Concerns regarding reliability persist, as 61% of organizations conducting pre-deployment evaluations reported at least one instance in the past year where an AI feature passed testing but subsequently failed in a customer-facing environment. Previous 2026 surveys indicated that only 5% to 13% of respondents fully trusted automated evaluations.
Data from earlier this year showed that 49% to 50% of deployments involved false positives, where tests incorrectly identified faulty features as functional. Currently, 20% of companies are working toward autonomous low-risk deployments, while 35% have opted out of such initiatives.
