MoEngage unveils Merlin AI Custom Agents, letting lifecycle marketers and CRM teams build and run their own workflow agents on MoEngage data within defined guardrails and coordinate with external AI systems via an open Model Context Protocol (MCP) server.

The Merlin expansion sits alongside MoEngage’s existing agentic platform, enabling enterprises to standardize on MoEngage while leveraging external AI capabilities through in-house and external agents.

An MCP Connector is being opened to integrate with external AI tools like Claude and ChatGPT, allowing external agents to access MoEngage data and coordinate with Merlin agents without bespoke integration work.

The platform supports both fully autonomous operation and manually reviewed operations, ensuring governance and safety as agents run continuously under marketer-defined guardrails.

This approach enables 24/7 agent activity and rapid experimentation, all while maintaining visibility and control through guardrails defined by marketers.

Merlin agents provide full transparency, exposing data sources, decisions, channels touched, and content sent, with audiences, channels, content rules, and budgets predefined before execution.

Customers like SoundCloud, Loblaws, Swiggy, and Domino’s helped shape Merlin, underscoring demand for transparent, controllable AI in marketing.

New capabilities include three Merlin AI agents: In-App Template Generator (creates responsive in-app templates), Flows Assist (builds a multi-stage journey canvas from an objective), and Campaign Insights Agent (answers performance questions in plain language with actionable recommendations).

MoEngage showcased Merlin AI Custom Agents and the MCP Connector at Salesforce Connections in Chicago on June 3–4, at booth #302.