COMMON POWERUp 2026

As enterprise AI shifts from experimentation to execution, Power Systems are becoming central to how organizations deploy and scale AI alongside mission‑critical workloads. This panel looks ahead at what’s next for AI on Power - examining emerging architectures, the impact of IBM Spyre, and how AI is shaping modernization efforts across applications and data. Panelists will share forward‑looking perspectives rooted in real customer environments, highlighting where change is happening, where focus matters most, and how Power customers can prepare for the next phase of enterprise AI without disrupting what already works.

Key Themes and Discussion Topics:
·       The shift to production‑scale AI on Power Systems
·       The role of IBM Spyre in advancing AI performance and efficiency
·       What agent‑driven and more autonomous AI means for core platforms
·       AI’s influence on application and data modernization
·       Extending existing IBM i and Power workloads with AI capabilities
·       Governance, security, and responsible AI considerations
·       Near‑term priorities for Power customers in the year ah

In this session, Tim will go over:

1. Why COBOL shops are facing a talent and maintenance crisis
2. How to use Claude to accelerate modernization
3. Flattening COBOL: copybooks, file defs, I/O expansion
4. Generating free-format SQLRPGLE from expanded source
5. Prompt tuning, context limits, hallucinations & lessons learned
6. Results ^ Q/

Your IBM i isn't going anywhere, but the systems around it keep growing. You have cloud platforms, third-party tools, and APIs that all need to talk to your core data and business logic. Getting them all to work together was already a challenge. Now, AI needs to fit in too.
In this session, we’ll show you how to bring AI to your current tech stack. We'll build a working AI application that connects to existing systems and data. We’ll connect to DB2 files, call RPG programs, and integrate with external APIs and cloud services. 
Along the way, we’ll cover the important decisions: data access controls, LLM selection, and AI governance so nothing touches your production environment without your say-so.
You'll leave with a clear picture of how AI connects to the data and systems you already have and how to build AI tools that fit into your business.

AI for IBM i is everywhere - AI you can trust in production is rare. IBM i teams don’t struggle because they can’t run analyzers or generate diagrams; they struggle because critical logic is scattered across code, data, and workflows - and “best‑guess” answers create validation loops, rework, and surprise blockers. This becomes even more visible as teams start introducing AI-generated code: without shared context and evidence, speed quickly turns into uncertainty. In this session, we’ll share a practical framework for trustworthy AI outcomes on long‑running systems - grounding, completeness, and traceability - and how to apply it without refactoring or disrupting existing workflows. We’ll also show how teams can use system-level understanding to validate and preserve knowledge as new (human- or AI-generated) changes enter production. The session closes with a live walkthrough of real IBM i questions to show what “good answers” look like in practice.

Most AI coding tools make you faster at writing code. CoderFlow does something different: it finishes the work — and it keeps working while you don't. In this webinar, we will walk through how CoderFlow, Profound Logic's enterprise agentic coding platform, executes the full engineering workflow autonomously. From compiling and testing to fixing failures and iterating until acceptance criteria are met, CoderFlow delivers verified, ready-to-commit outcomes instead of suggestions that still require you to drive every step.

Unlike copilot-style tools that pause and wait for your next prompt, CoderFlow runs continuously in the background — tackling queued tasks, resolving issues, and closing tickets without requiring a developer in the loop at every stage. Think of it less like a coding assistant and more like a tireless team member that works in parallel with your existing engineers, around the clock.

We will cover how autonomous agents work inside isolated containers, how the build-test-fix loop actually runs, and what work automation looks like in practice — including how teams are restructuring their sprint workflows, reducing toil, and reallocating engineering time to higher-value problems.

If you have been exploring AI coding tools and wondering whether there is something more capable for complex, real-world systems — or if you are ready to move from AI-assisted development to AI-automated delivery — this session is for you.

Learning Objectives:

• Understand the difference between AI assistance and AI automation — Learn how agentic coding platforms like CoderFlow move beyond suggestion-based copilots to autonomously complete end-to-end engineering tasks.
• See the build-test-fix loop in action — Discover how CoderFlow runs inside isolated containers, iterates on failures, and delivers verified, commit-ready code without manual intervention at each step.
• Explore real-world workflow transformation — Learn how engineering teams are restructuring sprints, reducing repetitive toil, and shifting developer focus from debugging AI output to reviewing completed work.
• Assess the impact on team capacity and throughput — Understand how running autonomous agents in parallel with your existing team can increase delivery speed without proportionally increasing headcount.
• Identify where agentic automation fits in your stack — Leave with a practical framework for evaluating whether CoderFlow is the right fit for your organization's complexity, systems, and engineering goals.

Ann and Ryan are both developers in the IBM i development lab. They both have access to IBM Bob and have been exploring ways to leverage AI development assistance as a part of their daily development activities. Ann and Ryan will share areas of success and even things that did not work well. This session is intended to help you gain insight into how IBM Bob can be a valuable development partner.

Curious what AI actually looks like in an IBM i shop? In this session, we’ll break down key AI concepts and patterns, then show working examples of AI running over IBM i data. You’ll see what happens behind the scenes, what guardrails are needed to keep things safe, and how these patterns can fit into your existing environment. You’ll leave with concrete examples, a mental model you can reuse with any LLM, and a clearer picture of how AI can work with IBM i instead of around it.

AI is quickly becoming a practical tool for improving how IBM i teams analyze, modernize and deliver applications. But knowing where it fits and how to apply it effectively is key. In this session, we will examine how AI can be used to interpret complex IBM i code, generate meaningful documentation, and assist in modernization initiatives. We will also cover how these capabilities can be integrated into existing development workflows to improve speed and consistency without introducing unnecessary risk. With real examples and practical guidance, you will learn how to move from experimentation to execution with AI in your IBM I environment.


Learning Objectives:

• Understand AI-driven techniques for analyzing and documenting IBM I applications.
• Learn how AI can assist with code transformation and modernization strategies.
• Explore integration of AI into existing IBM I development workflows.
• Get practical guidance on moving from AI exploration to implementation.

IBM Bob was first talked about back in October.  For then till GA a select group of folks from the IBM i community have take the opportunity to put Bob to work.  This is their stories.  How has Bob changed how they approach problems, projects and task, and how IBM Bob has made a difference.