Fusion Claw expands Oracle’s Agentic Applications portfolio to 75 total by adding 25 new, Claw-powered apps capable of deep research, computation, simulation, modeling, and continuous re-planning.

Governance, security, and audit capabilities are embedded, with enterprise-focused applications for Ledger automation in finance, Workforce Staffing in HR, Shipping Consolidation in supply chain, and Sales Territory Planning in sales.

Fusion Claw runs on Oracle Cloud Infrastructure, leveraging AI models from Google and OpenAI, with early access now and broader availability in the first week of October.

The core architecture splits reasoning and deterministic enterprise computation: LLMs handle reasoning and planning while deterministic systems perform calculations, data access, and transactional updates for exact, auditable results.

The approach emphasizes separating reasoning from execution and embedding governance to ensure reliability and auditability for large organizations.

A governance layer with role-based access, approvals, and end-to-end audit logs aligns with high-risk AI requirements and enables traceable decision trails.

Native transactional integration lets agents operate inside the organization’s transactional system and governance layer, reducing integration overhead and governance gaps.

Oracle argues for measuring AI by operational results rather than model usage, contending that separating reasoning from deterministic computation can cut costs on complex AI workloads.

Frontier models handle reasoning and planning, with deterministic enterprise computation executing at scale; an Outcome Trust Harness enforces controls per outcome run.

Buyers should assess whether agents run inside the governance layer or an integration layer, how tasks split between LLM inference and deterministic computation, audit trail strength for EU AI Act compliance, and clear error recovery paths for autonomous actions.

Automation levels range from assistive to fully governed auto-execution within delegated authority, with auditable Outcome Receipts documenting decisions, actions, and results.

The platform runs on Oracle Cloud Infrastructure, supports frontier models like Gemini and OpenAI, and plans to add more models, offering configurable autonomy levels from assisted to autonomous within delegated authority.