Oracle launches new AI Agent Studio in Fusion suite
The studio is designed for Oracle customers that already have most of their data situated in the Fusion suite. It includes new libraries and a trust and security framework.
Oracle, one of the first major players in the agentic AI arena, on Thursday launched a new platform for creating, extending, deploying and managing AI agents and agent teams across the enterprise.
Oracle AI Agent Studio for Fusion Applications is part of the Oracle Fusion Cloud Applications Suite, a group of SaaS applications that streamline and optimize business functions in industries like finance, human resources, supply chain and customer service.
The studio is free to Oracle customers and includes agent template libraries, agent team orchestration, agent extensibility and LLM choice. It also includes native Fusion integration, third-party system integration, a trust and security framework, and validation and testing tools.
Oracle AI Agent Studio is the latest in a string of agent-building tools on the market. Other platforms include Salesforce's AI Agent Builder and Microsoft Copilot Studio's Agent Builder. Oracle introduced its Oracle Cloud Infrastructure (OCI) Generative AI Agents service with retrieval-augmented generation in January 2024.
For Oracle customers
While it's not unique, AI Agent Studio serves Oracle customers whose data is already embedded in Oracle Fusion Cloud.
"If you have your most important and trusted data in Fusion, you want to build using the AI Agent Studio for Fusion," said Miranda Nash, group vice president for Oracle AI. "It's got just a native understanding of that critical data."
Oracle initially released OCI Generative AI Agents last year, with use cases for legal research, finance and customer service support.
If most of the critical data that you're going to expose to the model is embedded in Oracle apps, then you want to look at that.
Craig Le ClairAnalyst, Forrester Research
A problem with those agents was that they couldn't integrate well outside the Oracle environment, according to Craig Le Clair, an analyst at Forrester Research and author of Random Acts of Automation. Oracle AI Agent Studio differs from that by providing customers with an environment to customize their agents more, he said.
"This is a platform to build agents in a kind of bespoke way, which they didn't have before," Le Clair said. "It's a good thing for Oracle to be doing."
"If most of the critical data that you're going to expose to the model is embedded in Oracle apps, then you want to look at that," he continued.
Unifying all the data components in one environment is crucial for adding action components to agentic applications, he added.
Moreover, for Oracle customers, a tool like this makes sense since they mainly use Oracle products, according to R "Ray" Wang, CEO of Constellation Research.
"For Oracle customers, the ability to take all the Oracle portfolio and build agents is very powerful," Wang said.
This is also a progression for Oracle as a vendor that has invested in machine learning and embedding AI technology into its application platforms, Wang continued.
"It's good to be able to use agents as that interface and to be able to build agents to take advantage of a lot of the things that Oracle has done in the background," he said.
More governance
However, as with all agentic systems, enterprises are concerned about trust and governance, Le Clair said.
While Oracle addressed some of these issues with components like validation and testing tools that enable users to verify their AI-driven workflows, he added that there is still a need for explainability, bias control and ensuring that the agentic systems are not violating GDPR privacy regulations.
"You have these core governance elements as you move out toward more self-optimization and agents," Le Clair said. "We're not there yet. We're still developing the understanding or logistics to do this."
Esther Shittu is an Informa TechTarget news writer and podcast host covering artificial intelligence software and systems.
