How MoEngage built agentic customer engagement on Amazon Bedrock AgentCore
Machine Learning Blog
This article describes how MoEngage built Merlin AI, an agentic customer engagement platform on Amazon Bedrock AgentCore that lets marketing teams manage campaigns through conversational AI.
- Four agent modes: embedded agents in Merlin AI Studio, conversational agents for analytics, custom agents for continuous operations, and decisioning agents for individual customer actions
- MCP server exposes existing OpenAPI 3.0 APIs as agent-callable tools without rebuilding platform infrastructure
- AgentCore runtime provides multi-tenant isolation, JWT authentication, session management, and scales automatically
- Foundation models (Claude Sonnet 4.6, 5, 5.5) on Amazon Bedrock handle reasoning, tool planning, and content generation
- Tenant data isolation enforced through JWT validation, request headers, and isolated session environments
- Customer Glance achieved 50% faster campaign go-live and 38% uplift in push notification click-through rates
- Platform reports 60-70% reduction in manual effort and 20-30% improvement in conversion rates from goal-based decisioning
MoEngage demonstrates how software platforms can add an agentic layer by exposing existing APIs as MCP tools on managed infrastructure, enabling users to accomplish workflows through conversation rather than navigating multiple tools.
The AWS News Feed is currently looking for gold sponsors. If you want to support the AWS community and reach a large audience of AWS professionals, consider sponsoring the AWS News Feed.
Related articles
2026
2026
2026
2026
The AWS News Feed is currently looking for silver sponsors. If you want to support the AWS community and reach a large audience of AWS professionals, consider sponsoring the AWS News Feed.