Mixed-Agentic Estate: A Permanent Challenge for Organizations

Managing multiple AI agents across different platforms is a complex challenge that many organizations face today. According to IBM, this issue is not just an edge case but a common starting condition for most enterprises.

The problem arises when teams within the organization adopt different AI agent frameworks and tools without coordinating with each other. For example, one team might use agents built in a low-code studio, while another team uses LangChain-based agents for a specific workflow. Meanwhile, platform teams may build homegrown tools on top of MCP.

This mixed-agentic estate is not just a problem to be solved, but a permanent condition that requires ongoing management. IBM highlights three key reasons why this challenge is particularly hard:

Firstly, the pace at which new teams adopt new frameworks and tools is faster than any central function can standardize them. Secondly, most tools on the market only provide visibility into agent activity, not control over it. This means that governing agents you don't own or control is a much harder problem to solve.

Lastly, every cross-framework interaction creates a trust boundary crossing, which can lead to accountability issues when something goes wrong. To address this challenge, IBM emphasizes the importance of distinguishing between visibility and control capabilities and designing for agent governance from the start rather than bolting it on later.