The IBM i team wants to try something with the participants of COMMON Europe. We have a chance, at this conference, to collaborate with you on the kinds of things you'd like to see IBM i accomplish with AI in the future. Come to this session to participate as we use collaboration and design concepts to dream about what IBM i might do in the future.
STSM and Senior Business Architect - AI and Open Source, IBM
Experienced business architect from IBM and a specialist in the IBM i platform. My passion is enabling enterprises to adopt artificial intelligence and open source technologies.
In many IBM i environments, mission-critical business processes continue to rely on traditional 5250 green-screen interfaces, leading to challenges with knowledge transfer, skills shortages, and modernization pressures. This session explores a technical pattern for achieving secure, 100% reproducible AI-driven automation of real 5250-based workflows. By recording and analyzing actual operator interactions (keystrokes, screens, validations), these processes can be transformed into structured, documented representations—such as JSON-based microservice schemas—that capture exact input/output behaviors, validation rules, and flow logic without exposing underlying RPG/COBOL source code. These representations serve dual purposes: (1) enabling local AI models to execute the processes identically to a human operator (ensuring zero deviation and full auditability), and (2) acting as a living knowledge base for natural-language querying (e.g., "Explain the customer creation steps" or "Where do validation failures commonly occur?"). The approach leverages local AI deployments (no cloud data leakage), fine-grained security controls (e.g., JWT tokens and role-based access), and automated documentation generation to support onboarding, process optimization, and regression testing. Attendees will gain insights into the architectural building blocks, recording techniques, schema design principles, integration with IBM i security models, and practical considerations for maintaining fidelity in AI-assisted execution.This session provides technical guidance on bridging legacy IBM i processes with modern AI capabilities while preserving trust, compliance, and operational integrity in an era of workforce transitions.
Are you ready to adopt AI-based Assistants and Agentic AI tools on your IBM I? While these technologies are enticing and extremely powerful, there are limitations and pitfalls to factor into your AI strategy. During this session Carol Woodbury will discuss what the security considerations are that one must make when embarking on an AI project, how should one configure an Agent on IBM i and what guardrails you can put up to ensure it doesn’t go rogue, what auditing and logging is available to have visibility into AI activity on IBM i, how sessions using mapepire (e.g., MCP servers) manifest themselves in the audit journal, as well as what bad actors are doing with AI and how to protect your organization.
