Sandi MeyerAccording to new research from Enterprise Management Associates (EMA), enterprises don’t have an AI problem in network operations; they have a data problem.
Artificial intelligence has officially crossed the line from experiment to expectation in network operations.
EMA’s latest research, based on a survey of 458 IT professionals across North America and Europe, reveals that AI-driven NetOps is now viewed as critical to improving performance, strengthening security, and increasing operational efficiency. But here’s the uncomfortable reality: Only 35 percent of enterprises say they are completely successful in applying AI to network management. This gap between ambition and execution is more than a technology story; it’s a strategic one.
From AIOps Hype to Autonomous Expectations
Less than a decade ago, AIOps meant event correlation and anomaly detection. Today, large language models (LLMs), AI agents, and embedded intelligence are pushing vendors and enterprises toward autonomous operations.
EMA observes a decisive mindset shift:
- AI capabilities now influence vendor selection for nearly 60 percent of buyers
- Engineers increasingly view AI-driven insights as table stakes
- Strategic platforms are judged by their ability to automate confidently and explain decisions clearly
Skepticism has largely been replaced by expectation. But expectation doesn’t equal execution.
The Real Barrier Isn’t AI, It’s Data Readiness
When enterprises struggle with AI-driven NetOps, the root cause isn’t model sophistication. It’s foundational.
- Only 44 percent are fully confident in the quality of their network data
- Data silos persist across network and security domains
- Security and compliance risks limit AI automation
- Network complexity continues to expand
AI outcomes are only as strong as the telemetry data feeding them. If your AI is ingesting incomplete, sampled, delayed, or siloed data, you’re not building intelligent operations. You’re inferring rather than verifying.
Network Data: The New Competitive Advantage
EMA’s findings make something clear: High-fidelity network-derived telemetry data is emerging as the decisive differentiator in AI-driven operations.
Consider:
- 87 percent of enterprises are building IT and security data lakes
- For more than one-third, AI-driven operations is the primary driver
- Deep packet visibility, scalable telemetry ingestion, and cross-domain integration are becoming foundational
This reinforces a simple market reality: AI cannot compensate for blind spots in the network. Those without complete visibility will likely remain stuck in supervised AI. Organizations investing in enriched network telemetry, with deep packet visibility and full coverage across encrypted, East-West (lateral) and North-South traffic, are better positioned to:
- Trust AI-generated insights
- Reduce false positives
- Automate remediation safely
- Align NetOps and SecOps decisions
- Move toward autonomous operations
The Security Paradox
Security is both a catalyst and a constraint in AI-driven NetOps.
The research shows:
- 49 percent say AI-driven NetOps reduces security risk
- 46 percent actively use AI to address evolving threats, including AI-driven attacks
- Yet security and compliance concerns are cited as the #1 barrier limiting AI value
This paradox is critical. AI can dramatically improve threat detection, lateral movement analysis, and performance degradation tied to malicious activity. But without trusted data, automation introduces risk rather than reducing it.
Why This Matters
The competitive landscape is shifting. AI-driven network management is now operational, not experimental. It’s a buying criterion, an operational mandate, and increasingly, a board-level expectation.
EMA’s strategic takeaways for IT leaders are clear:
- AI is influencing vendor selection and long-term platform strategy
- Data readiness, not AI ambition, is the primary limiter of success
- Network visibility and data fidelity are prerequisites for scalable AI
- Cross-domain integration between network and security teams is essential
The organizations that win the next phase of NetOps won’t necessarily have the most advanced models. They will have the most complete, trustworthy, and actionable network-derived telemetry data.
Check out how the Gigamon Deep Observability Pipeline can help. And to explore the full research findings and detailed data from EMA’s January 2026 report, access the complete report.
CONTINUE THE DISCUSSION
People are talking about this in the Gigamon Community’s AI Exchange group.
Share your thoughts today
