Enterprises increasingly want AI agents that can reason over data spread across many AWS accounts without copying or centralizing it. Each team keeps its data in its own account for good reasons: clear ownership, scope isolation, and independent deployment lifecycles. But an agent that sees only one account’s data delivers limited value, and connecting it to distributed sources usually means replicating data or untangling cross-account AWS IAM policies. The goal is to let data stay where it already lives, in each line-of-business (LOB) account, ensuring that only the specific data a request needs flows out at query time.
In this post, we build a multi-account architecture using Amazon Bedrock #AgentCore Gateway and MCP (Model Context Protocol). Amazon Bedrock AgentCore is an agentic service for building, deploying, and operating highly effective agents securely at scale. A central platform account hosts the agent tier and LLM inference through Amazon Bedrock. LOB teams expose their data and tools as MCP servers, and the platform account’s AgentCore Gateway gives agents a single endpoint for tool discovery and invocation across registered LOBs.
The architecture follows a multi-account model with three layers: a central platform account, distributed LOB accounts, and AgentCore Gateway as the integration layer. The platform team runs the agent on AgentCore Runtime, a serverless, framework-agnostic environment featuring session isolation in dedicated microVMs, consumption-based pricing, and built-in authentication. The agent connects to the platform account's Gateway rather than to individual LOB #MCP servers, significantly simplifying connectivity.
LLM inference runs in the platform account through Amazon Bedrock. The platform team controls available foundation models (FMs), applies Amazon Bedrock Guardrails, and tracks costs through a single billing boundary, avoiding the overhead of managing model quotas across dozens of LOB accounts. As demand grows, organizations can distribute inference across several dedicated inference accounts, placing AgentCore Gateway in front as an Inference Gateway to route traffic across model providers and ensure enterprise-grade resilience.
[AgentUpdate Depth Analysis] The convergence of #AWS AgentCore Gateway and the Model Context Protocol (MCP) marks a pivotal step toward solving the "data gravity vs. global reasoning" dilemma in enterprise AI. By enabling a federated, zero-copy architecture, enterprises no longer need to compromise data governance for intelligence. MCP acts as the universal adapter for enterprise tools and data, while AgentCore Gateway operates like an API gateway designed for the Agent era—handling routing, session isolation, and cross-account authorization. This blueprint accelerates the deployment of secure, decentralized Multi-Agent Systems (MAS) in regulated industries, proving that the future of enterprise AI lies in federated execution.
