Oracle on Monday introduced AI-native developer tooling for its AI Agent Studio, giving Fusion Cloud Application customers the ability to build and deploy agentic applications using standard development tools rather than Oracle’s own low-code environment alone.
The new AI Studio Skill, announced via press release, supports Visual Studio Code, standard command-line interfaces, Git-based lifecycle management, and AI coding assistants including Claude Code. Developers can build Fusion Agentic Applications locally, validate and debug them, then push through CI/CD workflows. The resulting applications run inside Oracle’s governance framework, inheriting Fusion’s existing security, auditability, and approval controls.
What Ships
The update unifies three development tracks under one platform. Business users can design agentic applications through natural language in the existing no-code interface. Low-code builders get visual tooling. Pro-code developers now get the AI Studio Skill with IDE integration, local debugging, and version control. All three tracks produce applications that operate on Fusion business objects, workflows, policies, and logged actions.
Oracle is also expanding its AI Agent Marketplace and publishing reusable assets through a public GitHub repository. The company says all capabilities ship at no additional cost to Fusion Applications customers.
The Numbers
Oracle currently delivers more than 1,000 AI agents through Fusion Applications and launched 22 Fusion Agentic Applications earlier this year. The company claims over 80,000 certified Fusion experts who can now build on the new platform.
Enterprise Agent Development as Infrastructure Race
The announcement positions Oracle squarely against Anthropic’s Claude Code enterprise governance approach and OpenAI’s API-first agent ecosystem. Where those platforms require enterprises to bolt governance onto external frameworks, Oracle’s pitch is that agentic applications inherit existing Fusion permissions, approval chains, and audit trails by default.
The distinction matters for regulated industries and large enterprises already running Oracle ERP. A developer building a procurement agent or HR workflow agent gets the same compliance controls that govern their traditional Fusion applications, without stitching together separate identity, authorization, and logging systems.
The competitive axis for enterprise agent platforms is shifting from “can you build agents?” to “can you govern them at enterprise scale without a separate infrastructure layer?” Oracle’s bet is that its 80,000-strong certified expert base and existing Fusion security model give it a structural advantage over platforms that require enterprises to assemble governance from scratch.