Why the future of AI-powered customer interactions will be fueled by mainframe data (www.businessinsider.com)

IBM argues that unlocking mainframe data is critical to scaling agentic AI from experiments to business impact. The company highlights a new integrated approach—Salesforce Agentforce customers on IBM Z can use Salesforce Data 360 together with IBM watsonx and IBM DataGate (for watsonx and watsonx.data) to give AI agents Zero Copy access to mainframe data. Zero Copy lets agents read and use transactional and historical records in place (no costly ETL, middleware or API replication), delivering near real-time, compliant access to the 70% of global transactional workloads that live on mainframes. For the AI/ML community this matters because model quality and downstream automation are limited by data freshness and fidelity. Direct mainframe access reduces pipeline fragility, compliance risk, and integration cost, increasing the chances that agentic AI will deliver measurable value rather than be canceled for unclear ROI (Gartner forecast: >40% of agentic AI projects canceled by 2027). IBM cites tangible gains—customers who connected mainframe data were ~30% more likely to report cost savings and better predictions—and points to practical uses (fraud detection during onboarding, real-time loyalty lookups, airline call-triage). The takeaway: for high-stakes, time-sensitive, customer-facing tasks, in-place mainframe data access can materially improve autonomous decisioning, multi-agent orchestration, and end-to-end AI reliability.