IBM Centers AI Operating Model on Agentic Workloads

IBM's AI operating model addresses the challenge of running agentic workloads in production by combining four components: data, workloads, automation, and hybrid cloud. Ritika Gunnar, IBM's chief performance marketing officer, explained that enterprises must integrate AI with existing systems like SAP and Salesforce while maintaining governance, observability, and human oversight. Governed data with clear access permissions helps agents make better decisions. Evaluation loops assess agent performance against accuracy, cost, and timeliness, with business outcomes guiding how results are measured. IBM developed the model based on client experiences scaling AI across their organizations.