New Relic takes partnership tack for agentic AI
New Relic users say its collaborative approach to agentic AI, revamped AIOps tools and pricing model will make it easier to explore fresh forms of automation.
A bevy of updates from observability vendor New Relic match competitors in theme, but users say its approach to agentic AI and its pricing model set it apart.
New Relic made more than 15 updates to its Intelligent Observability Platform this week, including revamped AIOps features for incident management and predictive analytics; a refreshed toolset for business intelligence and service topology mapping; new data pipeline settings; FinOps support and an expanded list of agentic AI partners.
Many of the features unveiled by New Relic this week echo recent announcements from competitors Dynatrace and Datadog, which have emphasized support for new forms of AI-driven automation, including expansion into business observability. Agentic AI, in which multiple generative AI-driven software agents independently execute complex workflows, is a major buzzword for the entire IT industry in 2025.
But agentic AI is also where New Relic's approach differs most sharply from its rivals, according to New Relic's CEO. While Dynatrace and Datadog have integrated a broad set of functions into what began as application and IT monitoring software, including DevOps and security automation, New Relic will develop AI agents that primarily support such workflows in other tools.
"I asked a lot of our customers, 'Would you trust me with doing a software rollback?' And unanimously, they said, 'No, we don't trust you to do a software rollback," said Ashan Willy, who took over as CEO of the newly-private New Relic in late 2023. "I said, 'Would you trust me to stop a breach when you're being attacked?' Unanimously, they said, 'We trust you for a lot of things, but not that.'"
Instead, beginning with GitHub Copilot, Willy said, "We are going to be a system of intelligence for these other areas."
New Relic's AI agents will field queries from developers working in GitHub Copilot about an application's health and performance and return answers through the Copilot interface.
"That's the big difference with agents and just pure integration," Willy said. "We can receive that custom request and then query whatever we need to in our platform and say, 'Here's where it's going well. Here's where it's not going well.'"
New Relic features such as application debugging acquired with CodeStream will help inform debugging efforts for developers working in GitHub and similar tools such as GitLab and Atlassian's Bitbucket.
New Relic launched the agentic AI integration with GitHub last year and expects to make it generally available this spring. New Relic also previewed similar integrations this week with Amazon Q Business, Google Gemini and ServiceNow.
New Relic, Dynatrace, Datadog set strategic courses
Most observability vendors, including New Relic, Dynatrace and Datadog, now support agentic AI, and integrate with third-party tools. But their sales strategies and pricing models are notably different, said Stephen Elliot, an analyst at IDC.
For example, "On the security front, Dynatrace is not selling to the security operations center or security analysts but providing data to their core buyers to [share] observability data with security teams," Elliot said. "In contrast, DataDog is going after security budgets and selling to security teams."
The vendors' roadmaps have unfolded differently as they have created unified back ends for data collection and storage of various types of observability data, including logs. Datadog was the first to integrate application logs in 2018, while New Relic added application logs in 2019. After multiple attempts to integrate logs, Dynatrace introduced a new Grail data lakehouse back end in 2023, and made automated correlations between log and trace data in Grail generally available in 2024.
Pricing strategies for these vendors have also diverged over the last four years. New Relic moved to a unified back end for data and a pricing model that charges according to the amount of data ingested and the number of users with access to the platform, rather than the number of hosts on which its agent runs. Dynatrace added a cheaper pricing tier and data pipeline with fine-grained control over data ingest from various sources in 2024, but continues to charge per host for some services. Datadog also charges per host for each of its observability services, such as application performance monitoring.
New Relic's platform overhaul and pricing shift played a role in making it a private equity target, but one user said he expects its pricing approach to help going forward, especially as it rolls out agentic AI support.
"Normally, I would see stuff like this, and the first thing I'd be thinking about is, 'What's it going to cost me? What's the new license?'" said Steve Evans, senior vice president of engineering at Chegg, Inc., an education software maker in Santa Clara, Calif. "With how they do licensing now, we're just going to be able to start using it, and figure out what makes sense and what doesn't."
Among contrasts, some feature alignment
Despite their different overall strategies, feature updates from all these vendors over the last year have followed similar trends, including AI-informed responses to issues, predictive analytics for changes to infrastructure and applications, business intelligence and observability, and cloud cost management.
New Relic's Pathpoint business observability app previously allowed users to make connections between application services and business transactions, but a new app called Transaction 360 in preview this week automatically captures business transactions from beginning to end, Willy said. Transaction 360 is built on another new feature previewed this week called service architecture intelligence, which automatically creates a birds-eye view of applications and IT services.
"You manually construct it right now -- with Transaction 360 we automatically, without any tagging or anything, can see the entire transaction," Willy said. "Rather than map all those individual entities, we can just map a whole transaction right into Pathpoint, which makes it very easy to set up."
New Relic previously offered AIOps automation, but the new response intelligence feature previewed this week uses a retrieval-augmented generation mechanism and change intelligence agent to analyze information on prior incidents stored in third-party tools such as Atlassian's Confluence. From there, it provides insights on how to resolve the same problem if it happens again. AI-driven response intelligence will be available through agentic integrations with ServiceNow, Google Gemini, GitHub, Amazon Q Business, and more in the future, as well as New Relic's Incident Manager.
Predictions, another new service in preview this week, will become available throughout the New Relic platform as part of every query and chart, as well as the New Relic AI assistant. It applies machine learning models to predict future performance trends, set predictive alerts and generate forecasts on any chart.
Another customer said he's interested in New Relic's forthcoming agentic AI integration with ServiceNow, which his company uses for ITSM, incident and change management, and how Predictions might help analyze changes before they're made.
"Getting that real time flow happening between the two systems is going to allow us to quickly react, improve our mean time to recovery, mean time to respond, all of those kinds of operational metrics that are important for companies," said Corey Farrell, CTO at Pearson, an education software maker headquartered in London.
"I need to see how it's going to work in practice, but when changes get introduced into an environment, you want to be absolutely sure that they're not going to cause issues," he said. "If you look at resiliency statistics, a lot of companies struggle to make sure that what they put into production is going to work. So you see a lot of rollback mechanisms and time wasted in doing that."
Beth Pariseau, senior news writer for Informa TechTarget, is an award-winning veteran of IT journalism covering DevOps. Have a tip? Email her or reach out @PariseauTT.
