← Back

Wire · operational-macro

Enterprises Are Scrapping Agentic AI Pilots but Leaving Them Logged In

Published

10 September 2026

Topic

operational-macro

Sectors

CybersecurityAI Agents

Geography

United States

Source

Read at cybersecurity-insiders.com

Verified

Fusion42 · 10 September 2026 · Fusion42 review

A survey of 202 enterprises reveals that many have paused or scrapped agentic AI pilots but left the AI agents logged in with active credentials, creating significant security risks. Most enterprises overestimate their control and visibility over these agents, with many unable to audit or revoke agent access promptly.

This Wire brief sits within Fusion42's coverage of Cybersecurity and AI Agents.

◆ The Wire takeaway

Security founders must act immediately to help enterprises cut access for inactive AI agents. If you enable fast deprovisioning and real-time authorization checks, you're stepping into a critical opening few competitors see yet.

Coverage

1 source · 10 Sep 2026

Related on Wire

Topics

CybersecurityAI Agentsagentic-aisecurity-riskcredential-managementai-governanceenterprise-security