Oracle has expanded its Fusion Cloud Applications suite with new agentic AI capabilities designed to help healthcare organisations manage workforce requirements, clinical credentials, supply chains and other operational processes.
Speaking to Pulse+IT at the company’s Health and Life Sciences Summit in Orlando, Oracle vice president of healthcare product strategy Kristen Miles said the new capabilities would help bring together clinical and administrative systems and allow healthcare organisations to make decisions using information from across their operations.
Oracle Fusion Claw combines AI reasoning with deterministic computation to support complex enterprise tasks – including workforce planning, financial management and supply chain operations, with varying levels of human oversight.
Ms Miles said Fusion Claw was likely to have the biggest practical impact of Oracle’s recent Fusion developments.
“Fusion Claw is deep analytics in action – it unlocks a lot of value for customers, including the ability to have fully autonomous decision making based on your policies, bargaining agreements and union contracts.”
She said organisations would be able to introduce the technology incrementally, retaining human checks while determining how much autonomy they were prepared to give the applications.
Oracle has introduced 25 new Claw-powered applications across finance, human resources, supply chain and sales – bringing its broader portfolio to 75 Fusion Agentic Applications.
The applications run on Oracle Cloud Infrastructure (OCI) and can use frontier models including Google Gemini and OpenAI models, with support for additional models planned.
Rather than operating as standalone AI agents, Fusion Agentic Applications are embedded within Oracle Fusion Cloud Applications and can work with its business objects, workflows, policies, permissions, approval processes and transactional data.
Oracle said organisations can determine the level of authority given to the applications, ranging from AI assistance through to governed autonomous execution within defined parameters.
One of the new applications, Workforce Staffing, is designed to generate executable staffing plans based on factors including employee skills and accreditations, availability, schedules, labour rules, costs, coverage and service requirements.
It can model alternatives and continuously re-plan as circumstances change.
Other Claw applications include Ledger, for accounting reconciliation and exception investigation, Shipping Consolidation, for modelling logistics options and Account Territory Growth Plan.
Ms Miles said Oracle’s integration of its Fusion applications with Oracle Health systems was intended to connect workforce records with clinical operations.
“The nurse that is defined in human capital management is the same person that will be delivering care at the patient bedside through seamless integrations,” she said.
That could include checking whether credentials were current and matching clinicians to requirements such as specialist skills or language capabilities.
Oracle’s Fusion applications cover human capital management, finance, enterprise planning, supply chain and customer experience, with integrations into Oracle Health and Epic clinical systems.
Ms Miles said moving back-office systems to the cloud – with defined data fields and segments – had made those integrations more straightforward.
Building agentic applications
Fusion Claw follows Oracle’s earlier introduction of a new AI-native development environment within Oracle AI Agent Studio for Fusion Applications.
The environment supports no-code, low-code and pro-code development – allowing organisations and partners to build their own Fusion Agentic Applications using Oracle, partner and external agents.
Business users can use natural language through the Agentic Applications Builder, while developers can use the AI Studio Skill with tools including Visual Studio Code, command-line interfaces, Git, OpenAI Codex and Claude Code.
Applications are deployed natively within Fusion instead of through a separate agent runtime or external orchestration layer, allowing them to inherit Fusion’s security, approvals, governance and audit controls and execute against Fusion business objects and workflows.
The developer environment also supports local validation, debugging and CI/CD workflows.
Oracle said a public GitHub repository would provide templates, starter projects, sample applications, reusable assets and reference architectures.
Oracle has also been extending agentic functionality within Fusion Cloud Human Capital Management.
Recently announced capabilities include specialised agents for job architecture, talent profiles, role guide generation, learning and development and manager coaching.
The Intelligent Talent Profiles Agent, for example, can infer skills from connected work-system data to maintain employee profiles, while Skills and Learning Assignment Management allows administrators to use natural-language prompts to define employee groups, assign learning and monitor compliance.
Oracle has also introduced agentic course creation and autonomous learning-content authoring.
Australia and New Zealand
Ms Miles said several Australian healthcare organisations had been relatively early adopters of Fusion – while Tāmaki Health and Health New Zealand were among organisations using the applications in New Zealand.
She said Australian and New Zealand customers were also represented on Oracle advisory boards and participated in testing new functionality before release.
“We ask them not only to participate and present ideas as they’re testing features, but also before we even bring it to market, we ask them to access it, test it, kick the tyres,” she said.
Ms Miles said governance was becoming a major consideration as healthcare organisations looked to introduce more AI capabilities.
“The features that we’re delivering in Fusion are largely AI-based and they really need to have a plan for how they’re going to do that,” she said.
Healthcare organisations were also showing increasing interest in demand planning for supplies, an area that had traditionally relied heavily on institutional knowledge rather than dedicated demand planning capabilities.
Technical challenges included accessing source systems, mapping equivalent data across systems and dealing with data quality and completeness.
Ms Miles said Oracle was using mapping engines and AI to help align data, alongside seeded integrations and partners providing data cleansing services.
Another barrier was the variable cost of AI usage, which could be difficult for healthcare organisations accustomed to more predictable subscription-based arrangements.
Oracle said Fusion Claw includes built-in observability, return-on-investment measurement and safety controls.
Ms Miles said Oracle also used value calculation tools to compare the cost of AI usage with expected savings before capabilities were deployed, noting “We absolutely want to make sure clients know not to turn it on if it’s not going to save time.”
The author was a guest of Oracle.
Add Comment