As agent adoption scaled, we saw a common pattern emerge across enterprises, including our own sales organization: specialized agents deliver value, but without orchestration, users carry the cognitive load of choosing between them. At AWS Sales, this meant more than 20 domain-specific agents deployed across the global organization, with representatives context-switching between systems instead of focusing on customer conversations. In this post, we show you how we built Field Advisor on Amazon Bedrock AgentCore to solve this, the architecture decisions we made, and the measurable results that we’ve seen.

AWS sales representatives faced a significant challenge as AWS scaled AI adoption. With more than 20 domain-specific agents handling customer relationship management (CRM) operations, meeting scheduling, customer insights, product recommendations, and compliance checks, representatives needed to know which agent to invoke for each task. They also had to manage context across fragmented conversations and manually combine outputs from different systems. This overhead consumed time that could be spent understanding customer needs and delivering solutions.

The AWS Sales team chose Amazon Bedrock AgentCore because it provides the capabilities required for production agentic AI at scale:

These capabilities removed the need for custom infrastructure, so that the engineering team could focus on domain intelligence that improves customer outcomes rather than building foundational services. Field Advisor addresses the orchestration challenge by serving as a central layer that routes requests to specialized agents while maintaining a single conversational interface. Sales reps ask questions in natural language, and Field Advisor routes requests to the right agent or tool, maintains conversation context across multiple interactions, coordinates approvals for sensitive operations, and delivers unified responses. This ultimately enables faster, more informed responses to global sales needs.

Field Advisor serves as an internal conversational assistant that addresses six key workflows for AWS sales teams, each designed to maximize time spent on customer-facing activities:

Since launch, sales reps have submitted more than 120K prompts across all modalities, interactions that relieve the AWS Sales team to spend more time understanding customer needs rather than navigating internal systems. The human-in-the-loop component, which handles record creation and updates, saves large-scale sales representatives up to 2 hours per week, time redirected to customer conversations and strategic planning. The migration to Amazon Bedrock AgentCore delivered measurable improvements: 41 percent reduction in latency compared to the previous infrastructure, consolidation from seven separate AWS accounts to a single AgentCore Runtime, and the removal of custom-built systems for memory, observability, and authentication. The engineering team now focuses on product features that directly improve customer outcomes rather than infrastructure maintenance. Here’s what users say: