Multi Agent AI and Network Knowledge Graphs for Change — Ola Mabadeje, Cisco

Ola Mabadeje from Cisco's Outshift group discusses how AI can be used to reduce failures in production during change management. Outshift uses an incubation phase to understand customer problems, create prototypes, and deliver minimum viable products. Ola highlights the challenges of representing real-world networks with knowledge graphs due to the variety of vendors and data formats. The presented solution includes a natural language interface, a multi-agent system, and network knowledge graphs to create a digital twin of the production network. The system uses agents to interact with the knowledge graph, leveraging an open framework for building and integrating agents, and includes a demo of how the system works in a real-world scenario.