Oracle has introduced Fusion Clo, a runtime environment for managing AI agents that automate enterprise work. Some assess it gives customers tighter control over AI agents, while others warn dependence on Oracle could deepen.

The Register reported on Oct. 9 that Fusion Clo links with Oracle Cloud Infrastructure (OCI)-based Fusion applications to automate enterprise resource planning (ERP), finance and supply chain management tasks. It uses the latest AI models from Google, OpenAI and others. It plans to support models from other companies in the future.

Oracle CEO Mike Sicilia said, "Fusion Clo takes AI from assisting work to the execution stage." He said, "Customers can automate increasingly complex work in core operations."

Gartner assessed Oracle is not simply adding AI onto existing systems, but is reworking its application foundation to be agent-centric.

Processing large volumes of data with large language models (LLMs) can make costs rise to an unsustainable level. Gartner said Fusion Clo helps avoid that by having an agent make decisions and design tasks only once, while calculations are handled by a separate computing engine.

Forrester principal analyst Craig Le Clair said there are many trust issues that must be resolved to rely on agent systems.

Fusion Clo can be used only within Oracle products. The reality for companies is different. They build agents by mixing tools from multiple vendors, including Salesforce, Pega, UiPath and LangChain.

Le Clair said, "Platforms like Oracle are suitable for general tasks." He said, "For complex agents that need differentiation, it is better to use separate development tools or large cloud platforms."

The Register said Fusion Clo depends on Fusion applications and OCI, so companies that have already invested heavily in Oracle may welcome stronger management and cost savings, but it could be a burden for companies that use multiple vendors.