There’s networking for AI, and then there’s AI for networking, and Joe Vaccaro, senior vice president and general manager at Cisco Network Platform and ThousandEyes, breaks down the distinction on the latest Targeting AI podcast from TechTarget. According to Vaccaro, AI for networking assists network operations, while networking for AI builds resilient infrastructure for AI workloads and traffic. Vaccaro also outlines how agentic AI can help operate increasingly complex networks. He discusses the controls needed to make those agents trustworthy, from deterministic workflows and digital-twin validation to business-context guardrails and agent security.

In this episode, we discuss how:

• Cisco’s agentic loop moves through sensing, diagnosing, planning, validating, and acting.
• Cisco’s Deep Network Model and an ensemble of Cisco, small-language, and frontier models are presented as ways to apply network-specific knowledge and capabilities.
• Industrial and IoT environments bring physical AI workloads such as factory robots and vision systems, creating demand for focused models that can operate close to the edge in near real time.
• Cisco Cloud Control is a shared context layer for business intent, security policies, observability objectives and standard operating procedures.
• Agent security includes non-human identities with scoped roles and permissions, as well as protections against risks such as model jailbreaking.

To dive deeper into topics about AI and networking, check out TechTarget AI from Informa TechTarget, and please subscribe to our newsletter to keep up to date on the most important AI news.

To watch video clips from our podcast, subscribe to our YouTube channel, @EyeonTech.

References:

Cisco embeds agentic AI across customer journey

Cisco advances AI infrastructure services at Cisco Live

AI-native networking vs. traditional network automation