What You Missed at Oracle CloudWorld 2024 Part III - Revolutionizing Enterprise Development: AI, GenDev, and Oracle Fusion Applications

The evolution of Oracle’s AI capabilities was brought to life during the next session, where Juan Loaiza and Steve Miranda showcased groundbreaking innovations that are transforming enterprise development and applications.

Addressing Enterprise Development Challenges with GenDev

The next session was presented by Juan Loaiza, Executive Vice President of Mission Critical Database Technologies at Oracle. Juan talked about GenDev, security, low code on APEX, 23ai, Vector DB, and RAG. He first showed an application using GenDev so developers could get things to production faster during normal Enterprise Development. He explained that tools are not currently built for AI right now in enterprise development. It's like trying to race a sports car on roads that were built for horses in the 1800s. He showed that Enterprise App Development suffers from three main issues. The first are dependencies in the app. Next, there are large volumes of code that few people understand, and then you need the entire app to be both robust and enterprise-grade (security, availability, recoverability...etc.). We need AI-centric development where things can be built independently, trusted, and evolvable using human language. You also need security and speed.

Oracle Database 23ai: Innovation and Security at the Core

Enter Oracle's new product, GenDev, which stands for Generative Development, and was introduced with the Oracle database 23ai. You can use human language to build tables, indexes, entire applications, and security. A key feature of 23ai is JSON Relational Duality. A quote from Carl Olson, research VP, "Oracle's JSON Relational Duality is perhaps one of the most important Innovations in Information Science in 20 years." It's like a Microservice that provides JSON or Relational (SQL - PL/SQL) developers with transparent access to the same data. As described earlier, APEX delivers a way to use human language to create an application in just a few seconds.

Juan also talked about the 30+ in-database machine learning algorithms to predict the future, including classification, clustering, ranking, regression, feature extraction, attribute importance, anomaly detection, association, and survival analysis algorithms.

In-Database Machine Learning and AI Vector Technology

Juan then showed how the AI vector is just a series of numbers called dimensions, which are the semantic representation of a document, part of a document, an image, or a video. Oracle can generate, store, and compare different vectors without the data security issues associated with LLMs (Large Language Models). Still, it can also be used with the most popular LLMs (OpenAI, Cohere, Hugging Face, Llama, Google's Gemini, and VertexAI... etc.) if desired. He showed an APEX app that very quickly could find an apartment (when traveling) that looks exactly like the apartment I have in my home city by finding the vector distance between the image of my current apartment and the one I want to rent in another city.

Real-World Applications: PayPal and Vector Database Benefits

Juan then sat down with PayPal; they talked about how PayPal sees the Vector Database "as big" for its merchants to share and search things like images, videos, and so on with customers as needed. Customers can ask for exactly what they want in human language and search the Oracle Vector Database. PayPal also talked about Oracle's advantages, including true cache in 23ai, which gives data consistency with speed whether on one of the App Servers or in the Database. Juan then talked about how the Oracle Vector Search works seamlessly with other beneficial Oracle 23ai database features, including real application clusters (RAC - for availability) and parallel query for SQL transactions (PQ - for speed), all with great security as well as analytics. You can send Vector embeddings via Golden Gate in 23ai, and it all works transparently on Exadata, which makes vector searching events faster with the latest software (24.1) using AI Smart Scan.

Oracle AI in Fusion Applications: Transforming Business Processes

Steve Miranda, Oracle Executive Vice President for Applications Development, talked about Oracle AI inside Fusion Applications and how OracleAI is a part of all AI inside the fusion applications. Oracle uses AI in apps for Financials, Employees (HCM - Human Capital Management), Supply Chain (SCM), and so many more. In finance, 97% of the Oracle banking transactions are now automatically reconciled, 1.3 million employee hours are saved each year, the supply chain is now 70% faster for supply chain planning cycles, and in HR, two times the number of qualified applications are coming in now. In support, more than 1 million service requests (SRs) are routed to the correct support team, and some are handled by AI. AI was added into every part of Healthcare Solutions. Steve showed HR screens with analytics on the interviews, offers, applicants, and active offers as a draft, pending, accepted, or rejected. AI Assist can be used to answer questions in HR, such as providing a short description of candidate requirements and responsibilities. AI Assist can help with manager reviews, emails, and scheduling. Steve showed how resolved SRs were primarily being solved by automation. He also showed an example where 1.2 million SRs were solved by automation while only 253 needed to be solved by service reps (that's 5000x more solved by automation).

Steve also showed that there are 100+ Generative AI Features in Fusion Apps, such as a Chat Assistant, Work Order Assistant, Payment Assistant, and 50 plus AI Agents (a growing area for companies with AI capabilities).

Generative AI Features and Customer Success Stories

Steve then talked with Oracle customers Caesars Entertainment and DHL. Steve did a demo of an application where somebody put in an urgent requisition but sent one in Japanese. It was automatically entered into the English application, translated into English, and entered into the system. It was ready for payment online. This is the beauty of AI leveraged correctly!

AI and Productivity: Enhancing Workflows and Automation

Lastly, he showed how you can also use AI to answer emails, with a choice to click on something that says, “I want a longer answer or shorter answer.” I think this could make people lazy or more productive (depending on the person), possibly both. AI is being built into every aspect of every Oracle application.

Oracle’s advancements in AI and enterprise development tools, from GenDev to the integration of Generative AI features in Fusion Applications, underscore its commitment to innovation and business transformation. In the last section, we’ll explore how Oracle continues to empower its customers with cutting-edge tools for managing and optimizing their data and processes.