Where AI Earns Its Place in Mainframe Management

The mainframe is not going away. The Arcati Mainframe Navigator 2026 reinforces the continued importance of IBM Z to business-critical operations, while also showing an industry taking a pragmatic approach to AI. Rather than expecting AI to fundamentally disrupt the mainframe, organizations are looking at where it can solve specific operational challenges and create tangible value.

And that may be exactly the right approach.

Mainframe environments already contain enormous amounts of operational data and decades of specialist knowledge. AI creates an opportunity not to replace that expertise, but to make the knowledge and data behind the mainframe easier to access, understand, and act upon.

The central question for mainframe organizations is therefore: where and how can AI create meaningful value?

At SMT Data, this question guides how we integrate AI into ITBI, our SaaS platform for mainframe performance, capacity, and cost management. With more than 35 years of mainframe expertise behind the platform, our focus is not simply on adding AI functionality. It is on applying AI where it can help organizations get more value from their mainframe data and make specialist knowledge easier to access and act upon.

Our approach is iterative. We are integrating AI into ITBI step by step, focusing on areas where it can solve real problems while maintaining the trust, governance, and human expertise required in business-critical mainframe environments.

Where Can AI Add Value to Mainframe Management?

Mainframe environments already generate enormous amounts of operational data. The challenge is rarely a lack of data. It is making sense of it.

Capacity and performance teams need to understand what is driving consumption, where capacity is being wasted, when constraints might emerge, and what action should be taken. Answering these questions often requires specialist tools, detailed knowledge of the underlying data, and years of mainframe experience.

This is where we see AI becoming a valuable part of mainframe management.

Making Mainframe Knowledge More Accessible

One of the first opportunities is making existing mainframe expertise easier to access.

ITBI supports access through MCP (Model Context Protocol), enabling customers to connect their own AI tools to our mainframe capacity knowledge. Because MCP is an open standard, customers can work through AI tools that fit their own technology strategy rather than being tied to a specific AI provider.

Building on this foundation, Grace is our domain-specialized AI assistant for mainframe management, designed to make complex mainframe information easier to access and understand through natural language.

Instead of needing to know exactly which report, metric, or dataset to investigate, users can ask Grace questions such as:

“Are we at risk of hitting capacity limits before our next hardware refresh?”

or

“Where are we wasting capacity we are already paying for?”

The purpose is not to remove the need for mainframe expertise. Instead, AI can help make specialist knowledge and complex data accessible to a broader group of users — particularly valuable as experienced mainframe professionals retire and organizations face the challenge of transferring their knowledge to the next generation.

From Information to Recommendations

AI can also move mainframe management beyond answering questions to helping users identify what requires attention.

Traditional monitoring can tell us when a predefined threshold has been crossed. AI can add another layer by helping identify anomalies or emerging patterns, explaining why they matter, and providing guidance on what to investigate or do next.

In our conversations with mainframe organizations at SHARE, one point came through consistently: the organizations we spoke with were not looking for AI that puts proposed changes into production on its own. In business-critical environments, the stakes are simply too high.

What they are looking for is something more practical.

Consider a capacity manager arriving at work and being presented with the dashboards that actually matter — not only in technical terms, but also in business terms. What has changed? Which application, process, product, or department is behind it? What is the resulting financial impact in terms of MSU and TFP? Is increased capacity consumption driven by higher transaction volumes and business growth, or are technical issues behind it?

From there, AI can add the next layer: analyzing what has changed, providing context, and recommending possible actions. Those recommendations can then be evaluated by the people who know and own the environment, who ultimately decide what action to take.

The value lies in combining AI with trusted mainframe data and domain expertise, so recommendations are grounded in the customer’s actual environment rather than generated as generic AI responses.

From Insight to Foresight

AI and advanced analytics also have an important role to play in one of the core disciplines of mainframe management: forecasting.

Capacity management has always required organizations to look ahead. Teams need to understand not only what is happening today, but what increasing workloads, changing business demand, software pricing, or infrastructure decisions could mean in the future.

By bringing together historical mainframe data, domain expertise, and AI, ITBI can strengthen this forward-looking view — helping teams identify where capacity constraints are heading and providing earlier visibility into potential operational and financial consequences.

The real value is not prediction for its own sake. It is giving organizations more time and better information to make decisions before potential issues become actual problems.

How We Approach AI as an Organization

The opportunities are significant, but so are the pitfalls.

Serving banking, insurance, government, and other large enterprises running mission-critical infrastructure means AI cannot simply be introduced without considering how it affects security, compliance, data governance, cost, and accountability.

Five principles guide how we approach AI within ITBI:

Security and data ownership. Strict access controls ensure the right data reaches the right people — and no further. Customer data is never used to train external AI models.

Data quality is the foundation. AI is only as reliable as the data beneath it. Mainframe SMF data is complex and requires deep domain knowledge to normalize and interpret. AI can only provide trustworthy insight when the underlying data is equally trustworthy.

Cloud transparency. ITBI runs in the cloud, and so do the AI capabilities built around it. We are deliberate about where data is processed and stored. For regulated organizations, data residency is a compliance obligation, not an afterthought.

AI cost discipline. Running AI at scale has real costs. We design for efficiency, ensuring AI overhead is proportionate to the value delivered. In a world where every MIPS is scrutinized, introducing another source of opaque technology cost would make little sense.

Human judgment stays in the loop. When ITBI flags a capacity risk or recommends an action, a qualified person must evaluate and own that decision. We build AI to sharpen human judgment, not replace it. On a platform this business-critical, that is not a limitation. It is the responsible way to operate.

Practical Value, Responsible Adoption

AI has considerable potential in mainframe management. It can make specialist knowledge more accessible, help teams navigate complex data, identify important developments earlier, and ultimately support more proactive capacity and cost decisions.

But adding AI to a platform does not automatically create value.

For us, the important question is not how much AI we can add to ITBI. It is where AI can genuinely improve the way our customers understand and manage their mainframe environments.

That is the direction we are taking ITBI: combining AI with the trusted mainframe data and domain expertise already at the heart of the platform to make insight more accessible, more proactive, and more forward-looking.

At the same time, data integrity, security, transparency, cost efficiency, and human accountability remain fundamental.

The mainframe has earned its place at the heart of many of the world’s largest organizations through reliability and trust. As AI becomes part of how we manage it, those principles should remain just as important.