The rise of agentic AI in networking

Enterprises are becoming more comfortable with AI making network changes without human approval. However, companies still need to ensure that when AI agents make decisions, those decisions are not probabilistic.

The growing use of AI in networking is driven by the number of alerts flooding network operations, which have become too hard for humans to manage, according to Joe Vaccaro, senior vice president and general manager at Cisco Network Platform and ThousandEyes.

“If humans alone were simply responsible for resolving those alerts, you can’t hire enough people to go do that,” Vaccaro said on the latest Targeting AI podcast episode.

Cisco uses a four-stage framework to apply AI to network operations. With this approach, enterprises use AI to sense an issue, diagnose it using cross-domain telemetry, plan remediation and act.

Despite evidence that organizations are using AI in networking, a challenge remains: networks require deterministic, not probabilistic, outcomes.

Vaccaro said Cisco looks to build determinism into the action stage using predefined workflows rather than letting AI agents improvise changes.

“When you get to the stage of acting, when you get to actually applying the change across your network, you want to leverage a level of deterministic execution,” Vaccaro said.

About the Authors

News Writer, AI Business

Esther Shittu has covered AI technologies and industry trends since 2021. As co-host of the Targeting AI podcast, she talks with experts, thought leaders and practitioners exploring critical AI developments. Before AI Business, she wrote for SearchEnterpriseAI, the New York Daily News, Bklyner and the Brooklyn Daily Eagle. When she's not diving deep into the world of AI, she spends her time on passion projects and raising her three daughters.

Senior News Director, AI Business