More rope, fewer promises: what changed in nine months of running AI agents

Survey comparisons between December and September indicate that organizations are expanding AI agent autonomy while becoming more selective about where and how much they trust agents. Approval for every agent action declined from 38% to 30%, while intermediate governance and fully autonomous operation increased, though teams remained cautious about customer-facing communications and system-setting changes. Full confidence in agents operating independently fell from 37% to 32%, shifting largely toward partial trust, and urgency to change vendors for stronger agent capabilities declined even as overall willingness to consider switching remained high. Adoption is increasingly driven by leadership expectations rather than differentiated product ambitions, while platform priorities have shifted toward reliability, security, observability, and compliance over performance. These changes reflect organizations gaining production experience with agents, including mistakes and limited auditability, and focusing on governance, traceability, and safe deployment as agentic AI becomes a market expectation rather than a standalone differentiator.

Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.