"The talk covers how leading firms orchestrate agent workflows to handle complex tasks and deliver business value, with examples across telecom, finance, retail, and more. Importantly, it dives into practical considerations for scaling AI agents: establishing governance frameworks to ensure agents behave reliably and transparently, and implementing observability tools to monitor, debug, and optimize these dynamic systems in real time."
"Foundations with ADK, FastMCP & Tool Integration Google Cloud's Ivan Nardini and Annie Wang present the Google Agent Development Kit (ADK), a framework for building AI agents with modular workflows, memory, and tool use. They cover agent basics and demonstrate how to implement persistent memory and integrate external tools (like search or APIs). A key feature is FastMCP, an open-source Model Context Protocol implementation for secure agent-to-service connections. Attendees will build a working agent with long-term memory and tool-calling abilities using ADK and FastMCP."
Companies move from experiments to full-scale agent deployments, encountering immense opportunities and operational challenges. Leading firms orchestrate agent workflows to handle complex tasks and deliver business value across telecom, finance, retail, and other industries. Governance frameworks ensure agents behave reliably and transparently. Observability tools enable real-time monitoring, debugging, and optimization of dynamic agent systems. Modular frameworks such as the Google Agent Development Kit (ADK) provide persistent memory, modular workflows, and tool integration. FastMCP provides a secure Model Context Protocol for agent-to-service connections. Knowledge graphs combined with Retrieval-Augmented Generation (Graph RAG) enhance reasoning and reduce errors by translating business questions into graph schemas and converting unstructured data into agent-ready knowledge.
Read at Medium
Unable to calculate read time
Collection
[
|
...
]