Skip to content
hands-on lab

Build Multi-Agent Systems With Gemini Enterprise Agent Platform

Difficulty: Intermediate
Duration: Up to 1 hour
Students: 3
Rating: 5/5
Start lab
Get guided in a real environmentPractice with a step-by-step scenario in a real, provisioned environment.
Learn and validateUse validations to check your solutions every step of the way.
See resultsTrack your knowledge and monitor your progress.
About
Author

Description

In this hands-on lab, you will build a multi-agent IT support workflow with Gemini Enterprise Agent Platform and the Agent Development Kit (ADK). Gemini Enterprise Agent Platform provides a managed environment for visually designing, testing, and deploying AI agents. ADK complements the platform with a code-first framework for developing, orchestrating, and running agents.

You will create a ticket triage agent that collaborates with a specialist agent to identify the team responsible for an affected system. You will first design and test the system visually in Agent Studio, then export it to Python and run it locally with ADK. By comparing agent delegation with agent-as-a-tool composition, you will see how orchestration choices affect control flow and learn how to keep the main agent in control of a complete user interaction.

Learning objectives

Upon completion of this intermediate-level lab, you will be able to:

  • Create and configure a multi-agent system in Agent Studio
  • Define focused instructions and responsibilities for a root agent and a specialist agent
  • Export an Agent Studio project and run it locally with ADK
  • Explain how subagent delegation affects control flow in a multi-agent conversation
  • Convert a subagent into an agent tool

Intended audience

  • Developers building generative AI applications on Google Cloud
  • Cloud engineers exploring agent orchestration with Gemini and ADK
  • Solutions architects evaluating patterns for multi-agent systems

Prerequisites

The following prerequisites are beneficial for getting the most out of this lab:

  • Basic familiarity with Google Cloud
  • Basic Python knowledge
  • Familiarity with generative AI and prompt instructions

The following content can fulfill the prerequisites:

Environment before

Environment after

Covered topics

Artificial Intelligence
Google Cloud Platform
Artificial Intelligence for Google
Hands-on Lab UUID

Lab steps

0 of 5 steps completed.Use arrow keys to navigate between steps. Press Enter to go to a step if available.
  1. Signing In to the Google Cloud Console
  2. Building and Configuring the Ticket Triage Agents
  3. Authenticating the Google Cloud SDK (gcloud)
  4. Developing the Multi-Agent Systems With ADK
  5. Converting the Subagent to a Tool

Lab Rules