Oracle plans to train 1 million people on OpenAI tech for its new AI consulting arm

Oracle is setting up an AI consulting business powered by OpenAI technology, according to an announcement on its blog. Oracle Customer Success Services will lead the work, and the company has set a goal of one million people trained and certified on OpenAI products.

The new Oracle AI Services Practice is meant to help customers develop, deploy and scale AI. It combines OpenAI models, Oracle’s applications, data and cloud infrastructure, and the delivery teams of Oracle Customer Success Services (CSS). Oracle and OpenAI will train the CSS teams together, with OpenAI contributing its technical expertise. CSS runs customer engagements and delivery.

Four areas where Oracle will sell OpenAI-based work

Oracle describes the practice as planned to cover four areas, and the wording throughout is forward-looking.

• Fusion Cloud Applications. Customers will be able to build AI agents with Oracle AI Agent Studio for Fusion Applications using OpenAI models. Oracle Fusion Agentic Applications, with Oracle Fusion Claw, is meant for complex, long-running work. Customers can use built-in OpenAI models or bring their own large language model for supported OpenAI connections. Oracle experts will help with agent customization, testing for accuracy and reliability, and cost management through model selection.
• Application development on Oracle Cloud Infrastructure (OCI). OpenAI Codex and OpenAI APIs will be used alongside OCI Enterprise AI to build and run production software. Codex will handle development, integration, testing and evaluation. OCI Enterprise AI, including OCI Sandbox for isolated code execution, supplies the security and governance controls.
• Enterprise data. Oracle AI Data Platform, Oracle AI Database and Oracle Autonomous AI Lakehouse will be paired with a customer’s choice of foundation models, OpenAI’s among them, to build applications and agents that work across internal and external data.
• Security. Oracle AI Security Mesh is a managed service that uses OpenAI cyber models to find and fix vulnerabilities across OCI, other clouds and on-premises systems. The customer decides which systems may be tested and sets the boundaries, while Oracle runs safeguards meant to lower the risk of outages during testing.

Oracle CSS says it will cover the whole adoption cycle: Innovation Studios experts advise on strategy and use cases, specialists design and deploy solutions, and the team then helps monitor, secure and improve them after launch.

A seven-figure certification target via Oracle University

Oracle’s training plan, run through Oracle University with OpenAI’s collaboration, aims to put a million people through OpenAI-focused courses and certifications. The audience spans Oracle’s own consultants, its customers and its partners. The curriculum is to include role-based programs for technical and functional staff, hands-on Codex training, AI engineering courses focused on enterprise use cases, and Oracle University certifications.

Oracle frames the motive plainly: “Organizations are now shifting from AI experimentation to production.” The company says the training will help consultants and partners move customers from pilots to broader adoption, and strengthen customers’ own teams.

What changes for Oracle customers and partners

For existing Oracle customers, the practical change is a single vendor offering both the software and the people who configure OpenAI models inside it. Until now, customers wanting OpenAI models inside Fusion or on OCI typically relied on their own teams or outside integrators. Oracle is now positioning its own services unit for that work.

Related Oracle announcements give some context. Oracle’s separate deal to bring its business apps into ChatGPT was limited to US customers on the Oracle Marketplace side, so regional availability of this practice is worth watching.

For partners the picture is mixed. Oracle said days earlier that it would set aside $1 billion over four years for its partner ecosystem. A mass-certification program could equip those firms, but an Oracle-run services practice also competes with them for the same AI projects. The announcement does not say how work will be divided between CSS and partners.

What the announcement leaves unsaid

The post is a plan, not a product release. It gives no pricing, no start date for the practice, no timeline for the certification goal, and no list of countries. It names no customers and offers no results from engagements. The description of AI Security Mesh as finding and fixing vulnerabilities is Oracle’s own claim, and the post does not explain how access to OpenAI’s cyber models is controlled beyond the customer-defined scope.

It also does not say what OpenAI is paid or whether the arrangement is exclusive. The text stresses a choice of models for the data work and a bring-your-own option in Fusion, so customers are not necessarily tied to OpenAI, though the practice itself is built around its technology. Anyone considering the service should ask Oracle for the commercial terms, since none are published.