Written by Edward Calvesbert, Vice President of Product Management, IBM watsonx
The value of AI is determined by the data it uses. If AI agents don’t have real-time access to reliable data from modern mainframes designed for AI, they’re essentially working in a blind spot and risk using an outdated or incomplete view of the customer.
It can lead to erroneous conclusions, ineffective business decisions, mistakes and missed opportunities. Unlocking real customer data provides the context needed to move AI from “an interesting experiment” to a true growth engine, powering smarter strategies and personalized customer experiences that set businesses apart.
AI agents have great potential to automate business processes and interactions. They answer customer questions, analyze structured and unstructured data to derive new insights, and constantly optimize various schedules, resources, and constraints. But it’s not all about inference and automation. Gartner predicts that by 2027, more than 40% of agent AI projects will be canceled due to rising costs, unclear business value, or inadequate risk management. Maybe the problem isn’t the algorithm or LLM, but the data you’re using.
Mainframe data is an untapped advantage of AI
In industries such as banking, insurance, travel, and hospitality, valuable insights are often locked on the mainframe, including real-time customer interactions, financial transactions, and access to historical records. Mainframes process 70% of the world’s transactional workloads and are designed to integrate cutting-edge AI workloads into everyday processing.
However, this data is often overlooked as part of an organization’s AI strategy. why? Because until now, consuming mainframe data has been complex, expensive, and time-consuming. Integrations typically require middleware, APIs, or connectors, all of which come with high costs, pipeline vulnerabilities, and compliance risks. Zero-copy technology changes the game by allowing data to be shared and used on the fly without moving it. By eliminating the need for brute force data transfer, AI agents can access this critical corporate data in near real-time.
A new way forward with unified access
Data access rules are changing. Salesforce Agentforce customers deploying IBM Z can now access data without duplicating or costly transfers to AI agents. With IBM watsonx and Salesforce Data 360, agents get a broader, unified view of their customers and can act in real-time with increased security and trust.
Making mainframe data easily available to AI agents could lead to dramatic improvements. They are no longer just assistants. Act autonomously on accurate, up-to-date information across your enterprise with minimal human involvement. Gartner predicts that by 2028, at least 15% of daily business decisions will be made autonomously through agent AI, and 33% of enterprise software applications will include agent AI. Multi-agent orchestration allows you to automate entire workflows across domains.
Where data makes a difference
Not all AI use cases require mainframe data. But for high-stakes, time-sensitive, customer-facing tasks, this is a game-changer.
Consider the following questions as a measure of your use case:
- Does the workflow involve transactions, insurance policies, claims, or risk data?
- Is it customer-centric and time-critical?
- Are your agents making automated, mission-critical decisions based on insights from the latest data?
- Do agents need to collaborate across CRM, ERP, and other core systems?
If the answer to any of these is yes, then mainframe data with zero copies should be part of the equation.
How mainframe data impacts the real world
Banks can use fraud detection signals during customer onboarding. Call centers can quickly reveal loyalty information, warranty details, and service schedules. Airlines can now triage passenger calls more efficiently during weather disruptions, and customer service representatives are no longer forced to wait hours searching through multiple disconnected systems.
The results are measurable. According to IBM’s Salesforce study, customers who connected to mainframe data were nearly 30% more likely to report significant cost savings and more accurate AI predictions.
Agentic AI is only as effective as the data available. Siled data means mistakes and missed opportunities. With Salesforce Data 360’s zero-copy integration enabled by IBM DataGate for watsonx and watsonx.data, enterprises can benefit by providing AI agents with access to mainframe data, powering AI systems that deliver faster decision-making, sharper insights, and real business value to support customers.
See how IBM can help organizations break down data silos and power agent AI.
This post was created by IBM. Insider Studio.
