Dive Brief:
- One in four AI agents run unmonitored within enterprise environments, which can lead to operational risk as the technology autonomously alters code, configurations and infrastructure, according to New Relic’s 2026 Observability Forecast. The software company partnered with Enterprise Technology Research to survey 2,575 IT and engineering leaders for the report.
- As a result, the technology is fueling a surge in observability adoption. Eight in 10 organizations plan to deploy AI application observability within the next 12 months, the report found.
- CIOs are turning to AI-specific observability tools to determine whether the technology is working the way it should, New Relic CEO Ashan Willy told CIO Dive. “You’re going to have very business-specific outcomes that these agents are going to go after that you have to be able to observe,” he said.
Dive Insight:
AI workloads are driving demand beyond traditional observability tools, which are designed to measure the performance of cloud estates and other core systems. The shift is contributing to observability tool sprawl.
The number of observability tools increased in 2026 due to rising demand for AI-specific platforms, according to New Relic’s report. While organizations had reduced the number of observability tools from six in 2024 to 4.4 in 2025, the number is back up to an average of five in 2026.
Three in four enterprise IT organizations are using up to 12 tools across cloud infrastructure, network and service environments, according to Enterprise Management Associates research. Consolidating observability tools is a top priority for 62% of organizations, EMA found.
Vendors are responding as sprawl concerns mount. AWS on Tuesday launched Amazon CloudWatch Omni, an AI-powered observability console for AI agents. Earlier this year, Snowflake bought Observe, an AI-powered observability platform to increase insights for AI-driven enterprises, and Cisco added capabilities to Splunk for agent observability.
Unified platforms that absorb AI capabilities rather than adding new tools will allow enterprises to capture the benefits of AI without adding additional sprawl and complexity to the operational environment, New Relic’s report found.
“CIOs in particular have to always deal with actually making this stuff work,” Willy said. “It was very true in the early internet days, it was very true when CIOs were tasked with moving to the cloud and we’re in that period right now [with AI].”
While tool sprawl is a concern for businesses, New Relic also found that early high-impact outage costs fell from $76 million in 2025 to $73.5 million in 2026 due to faster incident detection and improved resolution times. Yet high-impact outage rates remain high, with more than one-third of organizations experiencing an outage at least once a week or more.
