Gemini Enterprise Sales Agent

1. 1

   Python

   Install

   Language for the ADK agent and FastAPI backend. Download the latest stable release.

2. 2

   Node.js

   Install

   Runtime for the workshop Next.js frontend.

3. 3

   uv

   Install

   Python package and environment manager. Replaces pip and virtualenv.

4. 4

   Google Antigravity or any IDE

   Install

   Code editor with agent support. Antigravity, Cursor, or VS Code all work well.

5. 5

   Google AI Studio

   Install

   Sign in with a Google account to access Gemini models and test prompts.

6. 6

   Google Cloud Agents CLI

   Install the official CLI and skills to build agents on Agent Platform.

   uvx google-agents-cli setupAgents CLI documentation

1. 7

   Clone the GitHub repository

   View GitHub repository

   Download the workshop code to your machine.

   git clone https://github.com/leonardoburbanov/gemini-enterprise-agent-platform-workshop.git

2. 8

   Enter the repository

   Move into the cloned project directory.

   cd gemini-enterprise-agent-platform-workshop

3. 9

   Get a Gemini API key

   Open AI Studio

   Create an API key in AI Studio with your Google account.

4. 10

   Create your first agent

   Scaffold a prototype ADK agent with your API key.

   uv run agents-cli create buildaiagent --agent adk --prototype --api-key <your-api-key-here>

1. 1

   Log in to Google Cloud

   Authenticate your GCP account interactively.

   agents-cli login -i

2. 2

   Create an agent step by step

   Interactive wizard to configure the agent in the cloud.

   agents-cli create -i

3. 3

   Deploy the agent

   Publish the agent to Vertex AI Agent Runtime.

   agents-cli deploy --deployment-target agent_runtime --project navbeai

Local playground of the production agent

Now let's look at the local playground of the agent that is in production.

We'll try it out to see what's happening in Google Cloud.

Gemini Enterprise Agent Platform

Production agent on Google Cloud

Traces and sessions

Cloud playground