Building an agent or an Agentic App is one thing. Testing every meaningful way it can execute, and continuing to test it as it changes, is a different challenge.
What this learning path is about
ATLAS (Agentic Testing and Lifecycle Automation Suite) is the testing framework for agents and Agentic Apps built with Oracle AI Agent Studio. It ships inside the AI Studio skill, runs on the same AI Studio CLI you build with, and you drive it in plain language through a coding assistant such as Codex or Claude. You describe the agent and the outcome you need; ATLAS works out what to test, prepares the data, runs the tests, judges the answers, and tells you what needs attention.
This learning path is for two groups of people who need to know an agent works, and keeps working as it changes:
- People who own or approve an agent. Business process owners, HR, finance and supply chain experts, and anyone who signs off before an agent goes live. You decide what the agent must get right and what it must never do. You don’t need to install anything.
- People who build and run the tests. Developers, functional consultants, and partners who build or adopt agents in Fusion with the AI Studio CLI.
It isn’t a training course and there are no quizzes. Each topic is a short, practical blog you can read on its own, built from real test runs.
Start here
If you own or approve an agent (no setup needed)
- The full ATLAS lifecycle: what happens at each stage, who does what, and when the evidence is enough to sign off.
- Why agents and Agentic Apps need ATLAS: why a few test questions aren’t enough, told through one HR agent.
- ATLAS: How it works alongside METRO: testing a change before release, and watching the agent once it’s live.
- The What this means for the business box at the top of Test the Worker Concierge agent, Review your test reports, Test an Agentic App, Improve cost and quality, and Cost guardrails. Each box says what was tested, what happened, and the decision it needed. You can stop reading there.
If you build or run the tests
Before you start, you’ll need the AI Studio CLI set up in VS Code with your coding assistant and a connection to your Fusion environment. If that’s new to you, start with the Learning Path for Fusion AI Agent Studio CLI, then come back here.
- New to agent testing? Start with blog 1 and read in order.
- Ready to run tests today? Go to blog 4, then blog 5.
- Inherited an agent someone else built? Read blog 7, then follow blog 8 step by step.
- Responsible for an Agentic App? Read blog 8 first, then blog 10.
- Running agents at volume? Read blogs 12 and 13.

The blogs
Each entry is tagged Everyone, Builders, or Builders, with a business summary (a short box at the top for owners and approvers).
Understand the problem and the tools
- Why agents and Agentic Apps need ATLAS · Everyone. Why a few input-and-output checks aren’t enough, what ATLAS is, and the sync plan that every ATLAS request starts from.
- ATLAS or METRO? Usually both. · Everyone. When to gate a change with ATLAS, and when to benchmark and monitor with METRO.
- Where ATLAS lives, and how to ask · Builders. Who does what, the AI Studio skill, what you need in place, and how the words in your request decide whether anything changes.
Run, read, and feed your tests
- Run your tests: a practical guide · Builders. Your first session, what happens during a run, and how to choose recorded or live data.
- Review your test reports · Builders, with a business summary. The reports you get, a review routine that works, how to read a failed report, and how far to trust the judge.
- Test data: how it’s recorded, and how to bring your own · Builders. Record once, replay anywhere, and keep it safe.
Test a real agent
- Build or pull: what ATLAS runs for you, and what it doesn’t · Builders. What runs on its own when you build or edit with your assistant, and why a pulled agent runs nothing until you ask.
- Test the Worker Concierge agent thoroughly · Builders, with a business summary. A real agent pulled from Fusion and tested end to end, including a green suite that never checked an answer and three tests only a person could write.
- What “thorough” means for each kind of node · Builders. A coverage checklist by node type, plus agents that pause for a person or a timer.
Test an Agentic App, keep it current, keep it efficient
- Test an Agentic App and its agents thoroughly · Builders, with a business summary. The Manager Coaching Workspace, its four panels, and the 32 workflow tests behind them.
- Keep tests in sync, and keep running them · Builders. What happens when the agent changes, regression on purpose, and bringing ATLAS into your CI/CD pipeline.
- Improve cost and quality through testing · Builders, with a business summary. Use the tests you already have to choose the right model for every step.
- Cost guardrails and ATLAS recommendations · Builders, with a business summary. Catch an agent that quietly got more expensive, and let ATLAS suggest what to improve next.
Reference
- The full ATLAS lifecycle: the whole journey on one page, with who does what at each stage.
- ATLAS glossary: every term used in the series, in plain language.
- Request library: ready-to-copy requests for your coding assistant, including a set for business reviewers.
- Command glossary: the CLI commands the assistant runs, grouped by job.
- Troubleshooting: what to do when a test or a run doesn’t behave.
Further reading
The ideas behind ATLAS are explained in the white paper Testing Agentic Applications: Why AI needs a new approach to software testing.
About the examples
Every example comes from a real test run in an Oracle demo environment. Screenshots are real; employee names, pay figures, and personal paths are masked.
