Managed-agent platforms package hosted, orchestrated AI agents for enterprise use: they provision models, manage state and connectors, apply governance and observability, and integrate agents into developer and business workflows. This topic examines how major providers — Anthropic (Claude), Google (Gemini/Vertex AI), Microsoft (GitHub Copilot and Azure integrations) and Amazon (CodeWhisperer evolving into Amazon Q Developer) — position their offerings across three categories: AI agent marketplaces, agent frameworks, and AI automation platforms. Relevance in 2026 stems from widespread enterprise demand for production-ready agents that balance capability with compliance. Organizations now expect multimodal models (e.g., Google Gemini), conversational and analytical assistants (Anthropic’s Claude family), and developer-focused agent features (GitHub Copilot, Amazon CodeWhisperer/Amazon Q Developer) to be available as managed services with secure connectors, role-based access, telemetry and lifecycle controls. Key platform distinctions are: marketplaces for third-party agent deployment and discovery; frameworks that let teams build, test and instrument agents; and automation platforms that embed agents into business processes and pipelines. When comparing vendors, focus areas include model access and multimodality, developer ergonomics (IDE and CI/CD integration), data isolation and governance, extensibility with enterprise systems, and operational observability. Practical enterprise choice hinges less on single-model performance and more on integration depth, compliance tooling, and lifecycle management — the capabilities that make managed agents safe, auditable and maintainable in production.