New Relic: 1 in 4 enterprise AI agents run unmonitored
2026-09-22
CIO Dive reports a New Relic study found that one in four AI agents in enterprise deployments run without monitoring, contributing to operational risk, and that AI adoption is obscuring system visibility, putting a spotlight on observability tooling.
Significance 3: A quantified figure on enterprise agent-monitoring gaps is a meaningful operational-lane adoption signal, consistent with the rubric's 'worth reading' tier.
Implications · machine-drafted, not owner judgment
Adds a quantified figure to the broader enterprise agent-governance gap already surfacing this month (echoing the earlier IDC/GuidePoint non-human-identity finding); reinforces that observability tooling, not model capability, is the current bottleneck to safe agent adoption.
- A named enterprise incident traced to an unmonitored agent
- Observability vendors report faster adoption of agent-monitoring products next quarter
Sources
- press 1 in 4 agents run unmonitored, giving way to operational risk retrieved 2026-09-23