What happened

The article notes that for the past three years the AI conversation in enterprise has centred on model output. The author, writing for SiliconANGLE, says the next three years will be about what systems are allowed to do. A Bloomberg Tech Event in Boston on 28 October, sponsored by Hewlett Packard Enterprise, will bring together chief information officers, chief security officers and data leaders to discuss deploying "agentic AI". HPE’s networking chief, Rami Rahim, told investors that AI is reshaping the technology stack, making the network more strategic. The piece outlines five areas CIOs should watch, including the shift from insight to action, the network becoming AI infrastructure, and security leaders setting the pace of autonomy.

Why it matters

Enterprises now have more AI-generated recommendations than they can act on. The real challenge is trusting an AI system to take action without a human approving each step. A mis-directed network policy change or a wrongly quarantined device could take a business offline. The article cites ServiceNow senior director Sajeev Nair, who described how consolidating telemetry reduced a manual workload from 2,000-3,000 hours a year to fewer than 60. That example shows the productivity gains possible when autonomous operations are correctly scoped. At the same time, the network must evolve from a simple pipe to strategic AI infrastructure, handling constant AI-generated traffic and acting as an identity and policy layer. Security teams, not just CIOs, will decide the speed of adoption because AI agents create non-human identities that traditional zero-trust models do not cover.

What this means for enterprise IT teams

IT leaders need a phased approach. First, define which actions can be fully automated, which need human sign-off and which remain manual. Second, measure outcomes from the user and business perspective rather than just infrastructure health. Third, ensure the network can support balanced upstream and downstream AI traffic and enforce identity-based policies for agents. Fourth, involve security chiefs early to design zero-trust extensions for AI identities. Finally, track success by tangible metrics such as engineer hours reclaimed, tickets avoided and faster time-to-value for new services, rather than by headcount reductions alone.

In the words of Rami Rahim, executive vice president, president and general manager of HPE Networking: "AI is reshaping the technology stack, making the network more strategic"

The bottom line

Moving from AI pilots to autonomous operations is not a switch-on event. It requires clear governance, network readiness and security alignment. Enterprises that treat autonomy as a gradual expansion of capability, measure outcomes that matter to the business and involve security leaders early will reap the promised efficiency gains while avoiding costly missteps.

Sources

  • siliconangle.com