Rapid expansion of the Fusion Applications AI strategy continues with custom AI

In less than two years, Fusion Applications have delivered hundreds of embedded generative AI use cases and embedded AI agents that are delivering productivity gains to Fusion Apps customers in production today. One month ago, we made Oracle AI Agent Studio for Fusion Applications available for our customers and partners to introduce AI agents into their workforce, across the enterprise. This is the same tool used by Oracle to deliver AI agents.

Embedded AI, Oracle’s unique value proposition

All the essentials, none of the complexity

All of this AI functionality has been delivered as embedded capabilities that are included with a Fusion Applications subscription at no extra cost. Oracle’s unique structural advantage enables the embedded AI strategy.

• Curated large language models (LLMs) are hosted in Oracle Cloud Infrastructure
• Because Oracle provides the infrastructure and data centers rather than requiring a third party, Oracle can also, uniquely, bear the cost of offering LLM inferencing via GPUs as part of its embedded AI for enterprise applications

By curating and hosting models in OCI, Oracle got to market quickly with embedded AI to reduce friction around AI adoption and help our customers put AI to work within existing enterprise applications workflows. As more companies deploy and iterate with AI agents, specialized needs have emerged among our customers and partners that go beyond off-the-shelf AI capabilities.

Now, in addition to embedded AI, Oracle will offer custom AI, giving customers more choice.

Unleashing limitless AI

For organizations with more complex AI needs, custom AI provides maximum flexibility, scalability, and control over AI-powered workflows. Custom AI is designed for AI power users who want to design, build, and deploy their own AI agent workflows and have those agents to be able to scale with their business without facing usage caps.

Key benefits of Oracle Fusion custom AI

• Maximum flexibility and control. Scale your business without limits and customize AI agents to meet the needs of rigorous AI governance processes.
• Unlimited usage. Unlock innovation with unlimited usage, empowering teams to scale AI-driven automation in the way that best suits their business.
• Deep customization. Create a lasting competitive edge with deep AI agent customization, hyper-personalizing AI by designing your own AI agents with Oracle Fusion.

Custom AI adds choice and control over your AI transformation

Limitless AI agents to scale your business beyond Oracle’s packaged capabilities

With custom AI, customers can introduce any number of fully custom agents and agent teams into production using the same AI Agent Studio tool that Oracle uses to configure and extend its AI agents. Customers can choose the models for their AI agents, without limits or keys. It also enables customers to deploy specialized multimodal models that can process images, audio, and video to power AI agents.

Custom AI will include OpenAI’s GPT-4 model to power the AI agents. Like the GPT offerings provided by other companies, Oracle’s GPT-4 models will be hosted in Azure. In the spirit of Oracle’s multicloud strategy, other high-performing models hosted in other clouds will be introduced over time to power Oracle AI agents.

Just like embedded AI capabilities enabled within AI Agent Studio, Oracle’s custom AI allows users to create agents using natural language; coding experience is not required.

Custom AI will be delivered as a separately priced offering within Oracle Fusion. Like other Fusion Applications products, custom AI will use predictable seat-based and employee-based pricing, following the pricing approach for core Fusion Applications.

Choosing between embedded AI and custom AI

We anticipate that most Fusion Applications customers will continue to take full advantage of the embedded AI capabilities included with their subscriptions, and Oracle will continue to deliver new embedded generative AI and embedded AI agents at pace. Some AI agent use cases will require custom AI due to the complexity of the use case or other specialized requirements that Oracle-built agents will not address.

With embedded AI, Oracle does all the work to maintain Oracle-built AI agents, all centered in Fusion Applications. Oracle curates and regularly updates models and agent configurations for high-quality results over time. With custom AI, teams that are capable of maintaining their own AI agents over time can now put those specialized AI agents into production to achieve limitless scale, reach, and automation results.

Embedded AI agents continue to be extensible using the embedded Oracle AI Agent Studio, which is able to connect to third-party systems, ingest customer-specific versions of documents, customize embedded prompts, and edit agent descriptions. Custom AI agents start from scratch without relying on Oracle-built agent templates. Both embedded and custom agents can be tested, deployed, and managed via AI Agent Studio.

Oracle embraces choice and openness because of the benefits to customers. AI agents are no different. Now, customers can choose embedded AI or custom AI or a combination of both to fully meet their enterprise automation needs with AI agents.