Research Finder
Find by Keyword
Can Mainframes Run Without the People Who Built Them?
BMC shifts focus from generic skills to capturing the unique institutional logic of the mainframe through governed, agentic AI workflows.
04/08/2026
Key Highlights
- BMC pivots to agentic AI to capture and scale proprietary institutional logic.
- New MCP and Agent Gateways centralize governance for autonomous actions.
- The zAdviser Enterprise Application Analysis automates the discovery of legacy application logic.
- Automated certificate management addresses the shrinking lifespan of security protocols.
The News
BMC has updated its Statement of Direction to move beyond basic generative AI assistants toward a coordinated ecosystem of autonomous agents. This strategy aims to capture decades of institutional knowledge from departing experts and embed it directly into the BMC AMI platform. The announcement introduces new diagnostic and security tools designed to help organizations manage complex mainframe environments with less reliance on specific individuals. Find out more by clicking here to read the press release.
Analyst Take
We see a significant shift in how BMC is positioning the future of the mainframe, moving away from a simple conversation about labor shortages and toward the much thornier problem of institutional memory. It is one thing to know how to write COBOL; it is quite another to know why a specific bank’s clearing system was architected in a particular way thirty years ago. In our view, the real threat to the enterprise is not a lack of general talent, but the evaporation of this specific, contextual knowledge as the original architects retire. BMC is attempting to solve this by building what we might call a "digital twin" of the organization’s collective expertise.
What Was Announced
The core of the announcement involves a multi-layered agentic architecture designed to move from insight to execution. Where other mainframe software vendors have taken the approach of waiting and watching AI play out, BMC is being more ambitious. Other vendors are simply rolling out a Model Context Protocol (MCP) server without a holistic approach to the wider architectural demands of leveraging the AI wave for mainframe application development and management.
This holistic approach from BMC includes an Intelligence Layer grounded in mainframe-specific telemetry and institutional data, a Coordination Layer for agent collaboration, and a Governance Layer for policy enforcement. A primary feature is the zAdviser Enterprise Application Analysis, which is architected to combine source code analysis with BMC AMI DevX telemetry. This tool aims to deliver a narrative intelligence report that highlights application risk and code complexity. Furthermore, BMC introduced the MCP Gateway and Agent Gateway, which are designed to serve as a shared, governed access layer for all AI-driven actions. Additionally, the new BMC AMI Digital Certificate Management solution is designed to automate the lifecycle of SSL/TLS certificates, preparing for the industry move to 47-day certificate lifespans.
We reckon the move toward an "Agent Gateway" is the most clever part of this strategy. Most enterprise AI today feels a bit like a collection of clever toys that don’t talk to one another. By forcing all AI agents to interact through a centralized gateway, BMC aims to deliver a transparent audit trail. This is a pragmatic way to handle the "governance" problem that keeps most CIOs awake at night. If an agent decides to modify a production workflow, the system programmer needs to know exactly why that happened and which policy allowed it. This architecture is designed to prevent agents from giving conflicting recommendations, which often leads to teams losing confidence and reverting to manual work.
We observe that BMC is focusing heavily on "Knowledge Hub" and "Knowledge Expert Chat." These features are architected to ingest runbooks, tickets, and log files. This is not just about making a better search engine; it is about turning decades of documented (and undocumented) fixes into a shared operational asset. Our analysis suggests that by embedding this intelligence directly into the workflow, BMC is trying to lower the "barrier to entry" for understanding bespoke legacy systems. The goal is to allow a junior engineer to act with the confidence of a thirty-year veteran by surfacing the right institutional context at the exact moment of decision.
The addition of zAdviser Enterprise Application Analysis is also a vital piece of the puzzle. It aims to deliver visibility into "knowledge concentration"—basically identifying which parts of the codebase are only understood by one or two people who might be leaving soon. We see this as a necessary step for succession planning. Rather than a manual, months-long audit, the AI is designed to surface these risks automatically. It is a bit like having a map of where the institutional skeletons are buried before they become a problem.
Finally, the move to automate digital certificate management shows a focus on "modernization in place." As security requirements become more aggressive, the manual management of certificates on the mainframe is becoming a spot of bother for many teams. This tool is designed to integrate with platforms like Venafi and Keyfactor, ensuring the mainframe is not an island when it comes to enterprise-wide security policies. We believe this focus on governed execution is what will separate the serious enterprise players from those just playing with chatbots.
Looking Ahead
The transition from AI assistance to orchestrated intelligence represents a fundamental change in mainframe management. Based on what we are observing, the industry is moving away from the idea that we can simply "train our way" out of the expertise gap. Instead, the focus has shifted to whether the software itself can inherit the institutional logic required to run the business. The key trend that we are going to be looking out for is how well these AI agents handle the edge cases that define mainframe operations.
Our perspective is that BMC is making a bold bet on "agentic" workflows as a cure for the loss of institutional memory. When you look at the market as a whole, this announcement aligns with recent observations from Bain and McKinsey, who suggest that the most successful AI implementations are those that target high-value, high-complexity domain knowledge. Comparing this to the wider market, BMC seems to be prioritizing a unified governance layer more aggressively than some of its peers.
Going forward, we are going to be closely monitoring how the company performs on the rollout of the MCP Gateway. The success of this strategy hinges on the ability to maintain a single point of control in an increasingly fragmented hybrid environment. HyperFRAME will be tracking how the company does in translating these architectural promises into actual reduced resolution times for its customers. If BMC can successfully capture the "how" and "why" of the enterprise, they will have turned the mainframe into a truly self-sustaining platform. This represents a tectonic shift in the value proposition of legacy hardware.
Steven Dickens | CEO HyperFRAME Research
Regarded as a luminary at the intersection of technology and business transformation, Steven Dickens is the CEO and Principal Analyst at HyperFRAME Research.
Ranked consistently among the Top 10 Analysts by AR Insights and a contributor to Forbes, Steven's expert perspectives are sought after by tier one media outlets such as The Wall Street Journal and CNBC, and he is a regular on TV networks including the Schwab Network and Bloomberg.