CIO.com Roundtable Explores Mainframe’s Role in AI-Driven Enterprise Transformation
Executive Discussion Highlights the Importance of Integrating Modernization, Hybrid Cloud with the Rise of Generative AI
By
Lane F. Cooper, Editorial Director, BizTechReports, Moderator, CIO.com
Kirk Chadrick, WW Leader - IBM Z DevX & Application Modernization
Last month, CIO.com hosted a virtual roundtable featuring top technology executives from industries including insurance, healthcare, banking, real estate, and manufacturing. The event, supported by experts from IBM, centered around the evolving role of mainframe platforms in supporting AI-enabled enterprises, particularly in the context of modernization and business transformation. As artificial intelligence (AI), and more specifically generative AI (GenAI), continues to reshape industries, enterprises are revisiting their mainframe systems to better align with this new wave of technological innovation.
AI and Mainframe: Current Focus on Proof of Concept
Not surprisingly, the roundtable discussion revealed that most companies are still in the early stages of exploring GenAI. Many participants indicated that their AI initiatives are in the proof-of-concept phase, with a strong focus on securing "early wins" with traditional AI and machine learning (ML) applications. These initial efforts are part of a broader strategy to qualify which data and applications can be augmented by AI before making long-term commitments. That said, several organizations highlighted their intent to develop AI/ML and GenAI strategies over the next 18 months to three years.
The Shift Toward Hybrid Cloud and On-Premises Repatriation
David Rossi, Cyber Security Architect, IBM
A key takeaway from the roundtable was the trend of "repatriation" — the movement of workloads from the public cloud back to on-premises systems. Since 2018, repatriation has grown by 20-30%, driven by challenges with cloud performance, cost escalations, and security concerns. Many organizations have recognized that while the cloud offers significant flexibility, fully migrating workloads can pose risks and inefficiencies. These concerns have been made more acute with the advent of GenAI and how organizations make use of large language models (LLMs) located in public cloud resources. It has prompted a strong line of discussion around whether data should be brought to AI resources “out there” or whether LLMs should be brought to the data “in here.” As a result, enterprises are increasingly adopting hybrid cloud strategies, combining on-premises mainframe systems with cloud-based solutions – in the process making deliberate decisions about where and how to manage enterprise GenAI activity.
In this context, IBM’s Kirk Chadrick, WW Leader, IBM Z DevX & Application Modernization, emphasized that mainframe technology remains critical for handling the computational demands of AI and large language models. “Mainframe technology has always been modern,” Chadrick stated, “but how it’s been used hasn’t always been up to date. We are working with customers to modernize their mainframe usage, particularly in hybrid cloud environments, to optimize AI utilization.”
Chadrick pointed to the role of mainframes in enabling real-time fraud detection using AI techniques, blending machine learning with large language models. This underscores how mainframes can continue to evolve, serving as the foundation for high-performance, AI-driven applications in enterprises
Addressing Shadow Data and AI Governance
David Rossi, a cyber security architect at IBM, addressed the growing issue of "shadow data," which refers to data that is scattered across legacy systems and cloud environments, often lacking proper governance. Rossi emphasized the heightened risks associated with unmanaged GenAI and data sprawl, warning that the lack of centralized data control could lead to increased costs and vulnerability to data breaches.
To counter this, Rossi suggested that organizations focus on centralizing and consolidating their data to reduce risks and improve visibility. In the context of GenAI, he noted that many companies will rely on pre-built models rather than developing models from scratch, but these models need to be rigorously vetted to avoid biases, legal risks, and inaccuracies. He also highlighted the importance of continuous model evaluation and human oversight to prevent "hallucinated" outputs, which occur when AI generates misleading or inaccurate information.
Looking Ahead: A Pragmatic Approach to AI and Mainframes
The CIO.com roundtable showcased a pragmatic approach to AI adoption, with participants emphasizing the need for balance between innovation and caution. As enterprises experiment with GenAI and other AI technologies, they are also recognizing the value of mainframe platforms in providing reliable, secure, and scalable computing power for their evolving AI workloads. The consensus is clear: hybrid cloud strategies, centralized data governance, and a cautious, phased approach to GenAI will likely define the future of AI-enabled enterprises.
For many organizations, mainframes are poised to remain a key component of this transformation, offering a modernized platform that is increasingly aligned with the demands of AI-driven business innovation.
###