Behind the Scenes – AI-Driven Modernization

This blog was written by Stephan La Rocca.

There is a great deal of momentum in software development right now. AI continues to advance with every release, changing not only how we write code, but also how we analyze, understand, and modernize existing applications. This development becomes particularly exciting where long-established enterprise applications meet modern, agent-based development environments.

With Oracle APEX 26.1 and the introduction of APEXlang, it becomes clear where things are heading: applications can increasingly be described in a structured, human-readable form that can also be used by AI agents. Instead of generating code in an uncontrolled way, application logic is described through clear specifications, rules, and traceable artifacts. This is precisely where skills, rules, and accurate contextual information become increasingly important.

For modernization projects, this is a crucial point. AI can only provide reliable support when it does not look at an application in isolated fragments, but understands its broader relationships. Individual Forms modules or isolated database schemas are not enough. What is needed is a complete view of the application, its dependencies, data flows, business logic, and technical structures.

Our experience shows that structured, fine-grained, and graph-oriented data from our PITSS.CON Application Data Cube provides an especially valuable foundation for AI-supported analysis and generation. It provides a more holistic overview of the application, captures the semi-structured meaning of individual clusters, and guides developers through implementation in a structured, dialogue-driven way. Based on the clear dependencies from the Application Data Cube, fragment generation speeds up developers' work.

This creates a major advantage, especially in complex modernization projects: AI agents can do more than provide superficial suggestions; they can work on the basis of reliable data. This improves the quality of analysis, accelerates implementation, and opens up new possibilities for an agentic development environment.

For the next version of PITSS.CON, we are therefore planning several interfaces that support its use in agentic development environments. This also includes an MCP server that enables structured application information to be made available specifically for AI-supported tools. The goal is not merely to expose individual technical artifacts, but to provide the complete application context.

Another important step is support for annotations in the Oracle Database. This allows additional semantic information to be generated, which in turn helps enrich customer applications with AI capabilities. This benefits not only developers, but also, in the long term, end users as applications become more intelligent, context-aware, and supportive.

AI is transforming software development at an unprecedented pace. The combination of PITSS.CON and AI opens new possibilities for analyzing, understanding, validating, and modernizing enterprise applications with greater speed, transparency, and confidence. By providing complete application context and precise structural insights, PITSS.CON enables AI to deliver more reliable results and accelerate modernization initiatives.