New EMA Research Reveals AI Adoption Gap in Network Operations

Why It Matters

The gap between AI adoption and success threatens ROI on multi‑billion‑dollar AI investments and could delay the shift toward autonomous network management.

Key Takeaways

• Only 35% report full AI NetOps success.
• Data quality identified as primary AI obstacle.
• 59% adopt vendor AI features; 52% train own models.
• Confidence in evaluating AI solutions below 40%.
• Network engineers now expect AI from strategic vendors.

Pulse Analysis

The surge of large‑language models and AI agents has transformed network operations from manual troubleshooting to predictive, self‑healing systems. Vendors are racing to embed generative AI into monitoring dashboards, anomaly detection, and capacity planning, positioning AI‑driven NetOps as the next frontier of IT automation. This wave of investment, however, is outpacing the readiness of many enterprises, creating a classic adoption‑success mismatch that EMA’s research now quantifies.

At the heart of the mismatch lies data hygiene. EMA found that only 44% of respondents trust the quality of their network data, and 39% feel confident evaluating AI solutions. Inconsistent telemetry, legacy proprietary formats, and incomplete documentation erode model accuracy, leading to false positives and missed incidents. Without a disciplined data‑management program—encompassing normalization, enrichment, and continuous validation—AI initiatives become costly experiments rather than strategic assets. The study also highlights a confidence gap: while 59% of organizations use vendor AI features, less than half feel assured of their ability to assess these tools objectively.

To bridge the gap, enterprises must adopt a data‑first roadmap. Establishing a unified data schema, investing in automated data‑quality pipelines, and integrating AI governance frameworks can lift confidence levels and improve success rates. Simultaneously, vendors should offer transparent model‑explainability and robust testing environments to help buyers evaluate solutions effectively. As network engineers grow more receptive to AI, the market will reward providers that combine sophisticated algorithms with clean, actionable data, accelerating the path toward truly autonomous network operations.