Episode
#578: How Cisco Is Using AI to Fix Networks
- Podcast
- David Bombal
- Published
- Jun 5, 2026
- Duration seconds
- 1351
- Processing state
not_requested- Canonical source
- https://soundcloud.com/davidbombal/578-how-cisco-is-using-ai-to
Actions
POST https://stenobird.com/v1/public/podcasts/david-bombal-5315180/episodes/578-how-cisco-is-using-ai-to-fix-networks/transcription-requests
Idempotently request low-priority transcript generation for this episode.GET https://stenobird.com/podcast/david-bombal-5315180/578-how-cisco-is-using-ai-to-fix-networks.md
Read the agent-friendly Markdown representation of this episode resource.
Summary
Cisco is bringing AI agents into network operations with Cisco Cloud Control, AI Canvas, and Agentic Ops. In this demo, David Bombal is joined by DJ Sampath (SVP and General Manager, AI Software and Platform) to look at how Cisco is using AI to simplify complex network troubleshooting, infrastructure management, agent security, and observability. Instead of jumping between multiple dashboards, tools, teams, and tickets, Cisco Cloud Control brings network, security, observability, and infrastructure context into one interface. The demo starts with a simple real-world problem: why can’t a phone connect to the network? From there, Cisco AI Canvas investigates the topology, calls the right agents, checks the wireless environment, moves into the firewall/security domain, and identifies the root cause: a site-to-site VPN tunnel issue caused by missing OSPF route exchange. You will see how Cisco is using MCP servers, topology agents, troubleshooting agents, firewall agents, and purpose-built models to help network engineers understand what is happening across Meraki, Catalyst, Firepower, Intersight, Splunk, Security, and other Cisco platforms. The video also covers the Unified Cisco Fabric app, which connects campus and data center environments with a firewall in between, plus Cisco’s agentic security app for monitoring and controlling AI agents in the enterprise. DJ also shows how Cisco is thinking about token usage, runaway agents, agent observability, Splunk, Galileo, policy enforcement, and secure AI adoption. The interview also covers Cisco’s work with OpenAI and Codex, including how frontier models, purpose-built Cisco models, MCP servers, APIs, Cisco data fabric, and real network data are being used to reduce hallucinations and make AI more useful for infrastructure te…