As enterprises embrace open-source technologies and AI-driven architectures, the challenge is no longer collecting telemetry data but turning it into actionable intelligence that drives resilience, performance and business value. In this Thought Leadership piece, David Cruddas, Senior Vice President and GM of EMEA at New Relic, explores why CIOs must move beyond data collection and embrace open, connected observability strategies to succeed in the intelligence era.
Open source is now mainstream – according to data from the Open Source Initiative, 96% of organisations have either ramped up or maintained their use of open-source software over the past year, with more than a quarter significantly accelerating their adoption. It’s clear the open-source ecosystem is now a mature community adopted by multinational corporations and AI-native startups alike.
For most enterprises, the decision to adopt open source is now driven by boardroom agendas to optimise costs and manage rapidly expanding digital infrastructure environments. McKinsey claims that technology leaders are embracing open-source tools as foundational to their technology stacks, citing clear advantages in high performance, ease of use and drastically lower implementation and maintenance costs when compared to proprietary software.
The shift to open-source has been especially profound in telemetry. OpenTelemetry (OTel) finally unshackled organisations from the restrictions imposed by proprietary software vendors in sharing data from metrics, logs and traces with other platforms. OTel gave power back to organisations to control how raw data flowed and was shared across multiple platforms, enabling them to eliminate blind spots across IT estates and use their data to make better and faster business decisions.
Data rich, insight poor
OTel may have given IT leaders access to richer data, but it created a new challenge: companies could now collect data from everywhere, but they lacked the tools to make sense of it. Especially for CIOs, who are navigating the uncharted waters of agentic workflows and hyper-distributed architectures, the sheer volume of data coming in began to impede their ability to use it effectively.
Enterprises started to drown in traces and logs and see their cloud storage bills and ingestion costs skyrocket, without seeing a meaningful drop in the time it took to resolve issues. Teams became data-rich but insight-poor because, although OTel streams carry inherent context, their tooling does not expose it.
This gap is widening rapidly with the introduction of GenAI and autonomous agents into the enterprise toolchain. Engineering teams do not work in isolation and neither do AI agents. When these agentic systems fail, they do not fail cleanly; they trigger a chain of dependencies across fragmented tools and hybrid infrastructure. Many legacy vendors are aware that enterprises are worried about this and are attempting to force enterprises back into closed-loop systems. They promise speed and simplicity, but only if you adopt their tightly controlled, proprietary environments to run and monitor your agentic workflows.
But this approach will seldom work for modern-day enterprises. Forcing open, modern architectures into closed-loop systems limits visibility, hampers engineering agility and creates operational blind spots. Walled gardens make it nearly impossible to apply consistent corporate governance, enforce risk policies or run reliable postmortems across distributed systems.
Closed platforms may offer the illusion of speed during a pilot phase, but they ultimately lead to severe fragmentation and costly failures at scale. Today, resilience doesn’t come from owning every piece of the stack; it comes from having the ability to connect them.
CIO priorities for the intelligence era
OpenTelemetry ultimately commoditised data ingestion and shifted the value of an observability platform from how much data it can gather to how intelligently it can reason across that data.
CIOs must completely re-evaluate their monitoring platforms in a standardised ecosystem. They must be capable of scaling alongside Agentic AI investments and treat open source as its native architecture. They must offer full, first-class support for OpenTelemetry and emerging AI communication standards like the Model Context Protocol.
Adopting OTel at your own pace
Transitioning to open standards must be made easy for modern enterprises. OTel is too risky if it requires a multi-million-dollar rip-and-replace initiative of existing platforms and tools. Platforms must offer a bridge which allows organisations to adopt OTel at their own pace while keeping systems online.
CIOs should look for platforms that offer native, hybrid agent compatibility. This allows engineering teams to instrument new microservices using OpenTelemetry APIs gradually, while ensuring these modern streams can coexist with legacy workloads under a single, unified view.
Managing data ingestion
Leaders must prioritise intelligent edge control to manage their data volume. OpenTelemetry makes data collection effortless, so unmanaged ingestion can quickly cause cloud storage and infrastructure bills to spiral out of control. The ideal platform must feature processing built directly into the telemetry distribution. In filtering out noise at the edge before it ever leaves the corporate network, enterprises can ensure they are only paying to store the actionable insights that matter to them.
Achieving connected intelligence
The risk of systems going down is greater than ever in modern architectures. A single outage occurring in a third-party service provider or SaaS subscription can lead to downtime that can cost organisations millions in lost revenue and reputational damage.
The companies that thrive in the next five years will not be those that collect the largest volume of uncompressed data points. The winners will be the organisations that leverage open architectures to achieve connected intelligence, turning raw telemetry into a distinct competitive advantage.
This strategy is already delivering undeniable business outcomes on a global scale. Data from IDC confirms that enterprises adopting an intelligent, unified approach to observability realise a 357% three-year ROI, achieving complete payback on their investment in just five months, and capturing an average of US$4.4 million in new revenue annually through optimised performance and reduced downtime.
Data strategy is in the hands of leadership
For organisations in EMEA, the evolution from data collection to intelligence is the defining challenge of the decade. OpenTelemetry has freed the enterprise from the constraints of the past, but it has also passed the responsibility of data strategy entirely to leadership.
To build a resilient enterprise capable of scaling AI responsibly, safely and cost-effectively, CIOs must reject the false comfort of walled gardens. By committing to an open and connected ecosystem, organisations can ensure their data flows seamlessly wherever practitioners need it. An open ecosystem is the only architecture that can achieve that.
