In a shift from conversational AI to autonomous workplace systems, enterprise infrastructure is becoming the bottleneck for deploying fleets of intelligent agents. During a conversation with Craig McLuckie, CEO of Stacklok, the Practical AI podcast explored how technologies like MCP (Model Context Protocol), Kubernetes, and emerging tools like ToolHive are enabling organizations to orchestrate multiple AI agents in production environments. The discussion centered on the infrastructure layer that transforms AI from individual user-facing applications to distributed systems operating behind the scenes across organizations.
The conversation highlighted critical challenges in building AI-native applications at enterprise scale, including identity management, agent orchestration, and system architecture. Rather than treating AI as a conversational interface, forward-thinking organizations are increasingly managing AI agents as coworkers with defined roles and responsibilities. This shift demands new approaches to infrastructure that can handle coordination, authentication, and monitoring at the scale of autonomous agent fleets. Stacklok's involvement in this space signals growing market recognition that the competitive advantage in enterprise AI lies not in models themselves, but in the systems that deploy and manage them.
Key Points
AI agents are transitioning from chatbot-like interfaces to autonomous coworkers performing backend tasks within enterprise environments
Infrastructure technologies including MCP, Kubernetes, and ToolHive are becoming essential for orchestrating and managing multiple AI agents in production
Enterprise AI adoption is increasingly constrained by system architecture challenges around identity management, coordination, and monitoring rather than model capability
Stacklok and similar vendors are positioning themselves to solve the orchestration layer for AI-native applications at scale