New Relic links AI observability to faster deployments

New Relic has published research that links use of its AI-based observability features with lower alert noise, faster issue closure and higher software deployment rates among its platform users.

The company said the findings came from aggregated, anonymised data from 6.6 million users of the New Relic Intelligent Observability Platform across 2025. The analysis compared accounts using New Relic AI with those that did not.

New Relic AI covers the platform's generative AI and AIOps functions. The company said these functions use a mix of machine learning, generative and deterministic methods.

Alert volumes

The report said engineers spend 33% of their time firefighting or addressing system disruptions. New Relic also reported that its customers handled 2.2 billion alert events during 2025. It said nearly 1 billion of those alerts originated in production environments.

New Relic said its AI-based observability features correlate signals into incidents. The company reported that accounts using New Relic AI recorded a correlation rate that was twice as high as non-AI accounts. It also said AI users generated 27% less alert noise than non-AI accounts.

The company framed the correlation and noise reduction figures as a measure of how much time engineers spend triaging alerts compared with time spent on other work.

Issue closure

New Relic used Mean Time to Close as the central metric for issue resolution. It defined MTTC as the average duration that issues remain open. The company said MTTC correlates with how quickly a system recovers from disruption.

According to the report, New Relic AI users resolved issues 25% faster on average than users without New Relic AI. New Relic also reported a wider gap during periods it described as high pressure.

The company highlighted one example month from the dataset. In May 2025, AI-enabled accounts averaged 26.75 minutes per issue. Non-AI accounts averaged 50.23 minutes per issue over the same period, according to the report.

Deployment rates

New Relic also connected reduced time spent on triage and "user toil" with code shipping frequency. It said non-AI users averaged 87 deployments per day during peak periods. It said AI-enabled teams reached up to 453 deployments per day during peak periods.

New Relic characterised that gap as a fivefold difference at peak. It also reported that teams using AI features shipped code about 80% more frequently on average than teams without the AI features during 2025.

The company did not provide a breakdown by region, sector or company size in the material it released. It also did not disclose how many of the 6.6 million users it counted as New Relic AI users versus non-AI users.

Product context

Observability tools collect and analyse data from software and infrastructure. Engineering teams use this data to identify incidents and performance problems. AIOps tools apply analytics and automation to operational data, including events and alerts.

New Relic positioned the report as evidence of a relationship between AIOps adoption and engineering outcomes. It described the relationship in terms of noise reduction, issue closure time and deployment frequency.

A New Relic executive linked the findings to the operational impact of AI systems in production environments.

"AI is bringing a level of complexity to modern software operations that is moving beyond human ability to manage. In tandem, AI is also helping solve the problem it created," said New Relic Head of AI Camden Swita. "The data from our user base tells the story of a new operational baseline for businesses. Those using AI-strengthened observability reduce noise and accelerate resolution, unlocking more time to focus on building new features rather than maintaining existing ones. Ultimately, this is a measure of an organisation's ability to respond to the market. When the operational tax is minimised, the speed of the entire company accelerates."

New Relic said it will continue publishing platform-based insights through its AI Impact Report series.