Building AI Agents with Google ADK
| Course Details | ||||||||
|---|---|---|---|---|---|---|---|---|
|
||||||||
| Languages | ||||||||
|
||||||||
| Outcomes | ||||||||
|
||||||||
| Price & Terms | ||||||||
|
||||||||
| Technical | ||||||||
|
||||||||
| Specification | ||||||||
|
||||||||
Learn to set up Google ADK and grow a Gemini support agent with knowledge-base tools, safety guardrails and specialist sub-agents in Python.
Tools take an AI agent beyond chatting, so it can look up an answer, keep to a support policy and pass a complex query to a specialist. Google’s Agent Development Kit supplies the classes for turning a Gemini model into an agent that does those things. You grow a customer-support assistant in stages, giving a simple greeting agent tools for knowledge-base lookups, guardrails that turn away unsafe or off-topic requests, and sub-agents that take over complex queries.
What You Will Achieve
- Set up an ADK project with working credentials and run an LlmAgent that replies in real time
- Give an agent a FunctionTool so it can pull answers from a structured knowledge base
- Add policy guardrails so an agent refuses unsafe or off-topic requests
- Build a multi-agent customer-support system in which specialist sub-agents take on separate tasks
Program Overview
Running a First Gemini Agent
Prepares the development environment and gets a minimal agent responding to a greeting.
- Installing Google ADK
- Credentials and configuration
- A minimal greeting agent
- Running the agent in a loop
Tools, Guardrails, and Multi-Agent Support
Extends the agent beyond conversation, then spreads a customer-support workload across several coordinated agents.
- Tool use and FAQ lookups from a knowledge base
- Safety guardrails and policy rules
- Single-agent versus multi-agent workflows
- A multi-agent customer-support system
Certificate Awarded (Paid)
Job Outlook
These skills apply to software work that puts language-model agents in front of users, such as customer-support assistants that look up answers, follow policies and route harder questions to other agents.
Good for: Developers who want a short, hands-on first build with Google’s Agent Development Kit, using a customer-support assistant as the running example.
Keep in mind: It builds one customer-support assistant across two short chapters, so tools, guardrails and multi-agent delegation each get a lesson or two in that single setting.
Bottom line: A compact pass through the kit’s core pieces that ends with a multi-agent support assistant using tools and refusing off-topic requests.
Frequently Asked Questions
What language model do the agents run on?
The agents run on Google Gemini. The first chapter installs and configures the Agent Development Kit, sets up credentials, and runs a basic Gemini agent that replies to a greeting in real time.
How does the agent answer support questions?
Through tool use. The lessons show how a tool lets an agent act beyond chatting, then connect it to a structured knowledge base so it can look up answers to frequently asked questions.
What happens when a customer query is too complex for one agent?
The later lessons contrast a single agent with a coordinated multi-agent system, then build a support setup where specialist sub-agents take on separate tasks and the main agent delegates the harder queries to them.
Do the videos have transcripts?
Yes. Each video has a live transcript that opens from a Show transcript control at its bottom left, and a course glossary sits in the resources panel on the right.
User Reviews
Write your review
You must be logged in to post a review.
Reviews can only be posted by registered learners. Log in or create a free account, then you will return here to post.
There are no reviews yet.