Splunk survey: AI agent rollout is outrunning monitoring and ownership
Original title: AI Agent Adoption Creates Risk as Oversight Struggles to Keep Pace, Splunk Finds
Of 2050 respondents, 38% lack clear ownership for monitoring AI and only 58% can effectively watch for hallucinations.
Cisco published the 2026 Splunk State of Observability on October 20, based on 2050 ITOps and engineering respondents. The report says organizations are deploying AI agents faster than monitoring and security controls, with responsibility split across engineering, infrastructure, web and security teams. While 85% said visibility into AI hallucinations is important, 58% said they can monitor for them effectively, and 42% said a different team owns observation. Separately, 86% said AI improved how fast they detect IT or security issues, while 14% said agents made issues harder to detect and resolve.
Unconfirmed: The percentages are from Cisco and Splunk’s survey, not an independent measurement.
- Mark Woods of Splunk said shared ownership of agent observability leaves it nobody’s job.
- 36% of respondents called the detection-speed improvement significant.
- Jeetu Patel of Cisco said security and observability data need a common platform.
- Some teams spend extra time checking whether an agent’s output is correct.
From the original article
2 min read at infosecurity-magazine.com
Rapid adoption of agentic AI comes with potential business benefits. But if done incorrectly it also carries risks around AI activity management and security
By Danny Palmer · infosecurity-magazine.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.