Recent Oracle presentations highlighted a clear shift in how organisations should approach AI. The focus is moving away from isolated pilots and experiments towards AI that is embedded directly into core business processes. From finance and HR to analytics and operations. This blog summarises the key insights. From a real-world cloud transformation case to the latest developments in Oracle’s AI database, AI agents and cloud infrastructure.
Cloud transformation in practice: Postcode Loterij
The Postcode Loterij provides a strong example of a large organisation successfully moving its core processes to the cloud. After 37 years, operating across five countries and generating annual revenues of €2.8 billion, further growth on legacy on-premise systems was no longer sustainable. The move to Oracle Fusion Cloud had one clear objective. Reduce the financial close from 25 days to just 5 days.
One of the most important success factors was simplicity. Rather than customising the software, the organisation adapted its business processes to Oracle’s standard. “Stick to the standard” became a guiding principle throughout the transformation.
Another key decision was to avoid a big-bang approach. The rollout started at headquarters, allowing processes to stabilise and lessons to be learned before expanding to other countries. This phased approach reduced risk and enabled continuous improvement during the rollout.
Oracle Database 26 AI and Generative Development
With the introduction of Oracle Database 26 AI, Oracle is positioning the database as a foundation for enterprise AI with long-term support. A core innovation is AI Vector Search. By introducing an AI vector data type, the database enables semantic search across documents, images and relational data. Not just based on keywords, but on meaning.
Oracle also introduced Generative Development. In this architecture, security is enforced at the data level rather than in the application layer. This allows AI-driven applications to be developed faster, while maintaining strong governance, security and compliance by design.
In addition, Oracle continues to invest in open standards. Support for Apache Iceberg enables interoperability with other data platforms such as Snowflake and Databricks, providing organisations with more flexibility in their data architecture.
From AI pilots to operational AI agents
A key insight shared during the sessions is that approximately 95 percent of AI pilots fail. Most fail because they lack access to secure enterprise data or do not deliver measurable business value.
Oracle’s response is to operationalise AI. Not as a standalone capability, but embedded directly into business workflows. Within Oracle Fusion Applications, more than 600 pre-built AI agents are already available. These agents support horizontal processes such as finance, HR and supply chain, as well as industry-specific use cases.
With the AI Agent Studio, organisations can also build, test and adapt their own agents. A core principle is human-in-the-loop. AI proposes actions, such as converting a quotation into a purchase request, while the user retains full control and final approval. There is also a clear distinction between different types of agents. Agents on the AI Data Platform focus on analytical insights. Agents within Fusion Applications focus on operational execution, such as invoice approvals or HR actions.
Infrastructure, multicloud and security
Oracle Cloud Infrastructure is designed to support demanding AI workloads. By using bare metal servers, OCI provides maximum isolation and performance, which is critical for data-intensive AI use cases. For governments and highly regulated industries, Oracle offers Dedicated Regions. This allows the full OCI stack to be deployed within the customer’s own data centre, while retaining the same cloud services and update model. Multicloud is another key development.
Oracle databases can now run directly within Microsoft Azure, Google Cloud and AWS data centres, with the same performance and pricing model as on OCI. This enables organisations to integrate Oracle capabilities into existing cloud strategies without compromise.
Costs, governance and data security
For Oracle Fusion customers, a significant part of the AI functionality is included by default. More than 400 AI features are available without additional licensing costs. Building custom agents in the AI Agent Studio is free in test environments. Production usage is priced based on the number of users or employees. An important security aspect is that the Large Language Models used operate entirely within the OCI infrastructure. Enterprise data does not leave the secured environment. This ensures full auditability and makes AI adoption viable in compliance-driven environments.
Conclusion: operationalising AI is the next step
The message is clear. The next phase of AI maturity is not about experimentation, but about integration. AI must be directly connected to operational data and business processes. By combining Oracle Fusion Applications, the AI Agent Studio and a secure cloud infrastructure, organisations gain a foundation to deploy AI at scale. Controlled, measurable and with clear business value. For teams looking ahead to 2026, the focus should be on turning AI into a practical capability that supports daily operations. That is where real impact is created.
Exploring how AI fits into your Oracle landscape? We support organisations in translating AI capabilities into practical, secure and measurable outcomes!
