A recent research from Enterprise Management Associates (EMA) reveals that network readiness is crucial for successful AI adoption. The report, "Readying Enterprise Networks for Artificial Intelligence," surveyed 269 North American IT professionals and highlights the challenges and strategies for preparing enterprise networks for AI workloads.
AI applications demand high-speed, low-latency, and lossless connectivity due to their unpredictable traffic patterns. This necessitates significant upgrades and re-architecting of both data centre networks and wide-area networks (WANs).
As AI workloads become distributed across public clouds, data centres, and the enterprise edge, the need for robust and adaptable network infrastructure becomes even more critical.
The EMA report identifies key business challenges, including security risks (39%), budget constraints (34%), and the difficulty of keeping pace with AI innovation (33%). These challenges reflect wider industry concerns about the complexities and costs associated with AI implementation. According to a 2024 Gartner report, AI security and risk management are top concerns for organisations.
Interestingly, 42% of companies have established AI centres of excellence to guide strategy across technical teams and business units. This indicates a proactive approach to AI governance and a recognition of the need for centralised expertise.
Furthermore, organisations that can automatically apply quality of service (QoS) or routing policies specific to AI-related traffic report greater success in preparing their networks. This highlights the importance of network automation and intelligent traffic management.
The findings underscore the need for IT teams to prioritise network optimisation as a foundational element of their AI strategies. Failure to address network readiness could undermine investments in AI technology. As Shamus McGillicuddy, VP of Research at EMA, notes, networks will "make or break" enterprise AI investments.
