A new report from New Relic reveals how artificial intelligence is fundamentally reshaping software operations, enabling engineering teams to resolve problems 25% faster and deploy code at dramatically higher rates when AI-powered monitoring tools are enabled.
The Alert Fatigue Crisis
According to the 2026 AI Impact Report, New Relic’s platform tracked 2.2 billion alert events across customer environments in 2025, with 821 million potential issues originating from production systems alone. This staggering volume highlights a critical challenge facing modern engineering teams: information overload.
The report, based on de-identified data from 6.6 million active users throughout 2025, found that approximately 33% of engineering time is currently spent firefighting or addressing system disruptions rather than building new features.
Cutting Through the Noise
The most striking finding involves “correlation rate”—how effectively software groups repetitive error messages into single, actionable incidents. Teams using New Relic’s AI capabilities achieved a 2X higher correlation rate compared to those without AI assistance.
This improvement directly addresses what the report calls “noisy alerts”—alerting rules with more than six triggering conditions that fire frequently but provide little useful context. AI-enabled accounts maintained an average noisy-alert rate of 46%, compared to 63% for non-AI accounts, with some non-AI environments exceeding 70%.
Faster Problem Resolution
The speed benefits are substantial. During peak performance in May 2025, AI-enabled accounts resolved issues in an average of 26.75 minutes, compared to 50.23 minutes for accounts without AI—roughly a 50% reduction. This 23-minute advantage per issue compounds quickly across hundreds or thousands of daily incidents.
Engineers using New Relic’s AI chat features submitted 1.63 million prompts throughout the year, primarily for query execution, investigation assistance, and documentation retrieval. The report suggests this reduces the manual “detective work” phase of troubleshooting, providing immediate context and suggesting relevant queries.
The Deployment Dividend
Perhaps most significant for business leaders: the productivity gains translate directly into shipping velocity. At its 2025 peak, AI-enabled accounts averaged 452.99 deployments per day, compared to just 87.04 for non-AI users—a 5X increase.
Across the full year, AI-enabled teams maintained deployment rates approximately 80% higher than their peers. The report frames this as an “innovation dividend”—cognitive capacity previously spent on operational toil now redirected toward building new features and responding to market demands.
A New Operating Standard
The New Relic report identifies what it calls a “compounding effect” or “engineering velocity flywheel.” Better signal filtering leads to faster incident resolution, which frees engineering hours for more frequent, confident deployments. The data shows these improvements sustained consistently across multiple months and metrics.
The findings reflect real-world conditions across diverse architectures, team maturity levels, and deployment practices. While operational performance depends on many variables, the consistency and magnitude of the differences between AI-enabled and non-AI accounts suggest quantifiable benefits from AI-enhanced observability.
For organizations still operating without AI-strengthened monitoring, the report’s conclusion is pointed: the question isn’t whether AI adds value to software operations—it’s whether companies can afford to operate without it when competitors are shipping code five times faster.
