Foundations
What Fusion AI is made of, how one run travels through it, and what it costs.
Oracle Fusion AI · interactive course
Every lesson is a diagram that moves: watch a request travel through workflow nodes, tools, approvals and policies, see what breaks, and see what fixes it. Workflow agents, tools, agentic apps, Fusion Claw, credentials, METRO and ATLAS, with notes from real builds.
423 animated steps. In a lesson, use ← → to step, R to replay, Space to pause.
What Fusion AI is made of, how one run travels through it, and what it costs.
The graph around the model: nodes, state, branches, loops and triggers.
What agents can call, whose identity each call carries, and how to keep writes apart.
Workspaces where several workflow agents share one process: panels, actions, chat.
Identity, credentials, policies, approvals and gates: keeping agents inside the lines.
You define the outcome; a planner, deterministic compute and a trust harness do the work.
METRO and ATLAS: seeing what agents did, scoring how well, and testing before you ship.
Drafts, publishing, promotion between pods and building from code.
How Fusion agents connect to everything else: OIC, MCP and A2A, FSCR, APEX, and Oracle's data-side agents.
Independent course, not an Oracle product. Based on Oracle documentation, release notes and hands-on builds; check Oracle's documentation for your release.