HN792: Understanding Agentic AI for Network Operations (Sponsored)

The podcast explores agentic AI and its potential to revolutionize network operations. Juniper Network's Chief AI Officer, Bob Friday, defines agentic AI as the next step in automation, using graphs of interconnected agents leveraging large language models (LLMs) for non-linear, non-deterministic problem-solving. He emphasizes the importance of reflection, tool use, planning, and reasoning in AI agents. The discussion covers how these agents, acting as specialists, can troubleshoot networks by accessing data from various sources, including databases and APIs. Friday also addresses concerns about hallucinations and the need for trust, suggesting that AI should perform on par with entry-level human employees. He touches upon security, data privacy, and the balance between public and private models.