02 / IT · Analysis
IBM and Microsoft unveil enterprise AI agent platform
Imagine Enterprise centres on adapting pre-built assets to company workflows, while its customer example focuses on technology exploration.

IBM announced that Imagine Enterprise, its new platform with Microsoft, would combine more than 6,000 pre-built AI agents and related assets [collection time not provided]. The figure was IBM's own tally; it did not disclose separate counts for agents and other assets.
The companies presented Imagine Enterprise as a platform for building, running and governing AI agents for business processes. Their approach centred on adapting existing capabilities to each company's data and workflows.
The companies also described connections to existing systems, with security, access permissions and controls built into the design. That made integration and operation central to the proposal, alongside the agents themselves.
Adapting pre-built assets to business workflows
IBM said companies could tailor pre-built agents to their own needs and connect them to existing data and business processes. The starting point was a set of reusable capabilities that customers would adapt.
IBM described a business combining human expertise with AI agents as its goal for customers. It said the initial industry frameworks covered financial services, insurance and retail.
Microsoft's technology foundation
The companies said Microsoft would provide the AI, data and security foundation. The design also included enterprise identity and control functions to support agents operating across a business.
The companies said qualifying projects would receive engineering support from both partners. They described that support as helping customers move from experimentation into production.
Customer exploration and operating choices
IBM's announcement presented SEI, a financial technology, operations and asset management services provider, as exploring agent technologies. It did not disclose performance figures for that collaboration.
IBM also offered deployment and operation on a customer's behalf, joint development, and self-management options. Its proposal therefore left customers with a choice about who would build and run the platform as they moved towards operational use.