Oracle announced Oracle Fusion Claw on September 29, 2026, describing it as a governed agentic execution runtime for Oracle Fusion Agentic Applications.
According to the Oracle announcement, Fusion Claw is designed for larger-scale, longer-running work, including deep research, computation, simulation, modeling, and continuous re-planning. It arrives with 25 new Claw-powered applications, which Oracle says bring the Agentic Applications portfolio to 75.
The release continues a fast build-out. Oracle launched Fusion Agentic Applications in March 2026 with 22 applications across finance, HR, supply chain, and customer experience, and in April it added 25 more across the Fusion Cloud portfolio.
Those earlier releases established what agentic applications do inside Fusion. Fusion Claw addresses how they handle more complex work and which controls apply when they do.
Fusion Claw Separates AI Reasoning From Enterprise Computation
Oracle calls each Claw run a Claw Outcome. A run starts with a frontier model that reasons, plans, and adapts as conditions change. Fusion Claw then hands execution to deterministic enterprise computation, which processes high volumes of transactions precisely and repeatably.
Oracle says this division gives customers a potential economic advantage. AI reasoning is applied only where judgment is needed, and more efficient computation handles the volume. The runtime runs on Oracle Cloud Infrastructure and uses frontier models including Gemini and OpenAI, and Oracle plans to support more models over time.
The launch applications show the pattern in practice. Ledger works across large volumes of accounting entries to reconcile information and investigate exceptions ahead of close. Shipping Consolidation models alternatives across timing, capacity, cost, and service commitments, and then carries the best-supported plan into execution when authorized. Workforce Staffing and Account Territory Growth Plan apply the same approach to HR scheduling and sales coverage.
Governance Is Set Before Each Run and Recorded After It
Fusion Claw governs autonomous work through three components.
The Enterprise Operating Envelope captures an organization’s objectives, standard operating procedures, policies, permissions, risk thresholds, decision rights, approval requirements, and escalation boundaries. An Outcome Trust Harness applies that envelope to each run and controls identity, capabilities, data, and actions. When a run finishes, an Outcome Receipt records the authority applied, the evidence used, the decisions made, the transactions executed, and the result.
Customers also decide how much automation each process gets. The range runs from quick assistance to governed full-auto execution within authority they explicitly delegate. The applications sit within Oracle AI Agent Studio for Fusion Applications. There, the Agentic Applications Builder lets organizations create and manage agentic applications with built-in observability, ROI measurement, and safety controls.
What This Means for ERP Insiders
Well-documented processes are ready to automate first. The operating envelope draws on the policies, decision rights, and approval rules that ERP teams already maintain. Teams that bring that documentation up to date give themselves a faster, more confident path to higher automation levels.
Outcome Receipts give audit teams a clear record to work from. Each run leaves an account of what was decided and on what evidence. Bringing in controllers and internal audit early to agree how receipts fit existing control reviews makes it easier to expand automation with their support.
Graduated automation supports a staged rollout. ERP teams can start with assistance on high-volume work such as close exceptions and track results through AI Agent Studio’s ROI measurement. They can then move toward full auto as confidence grows, building evidence for each step before taking the next.
