B2C Lifecycle and CRM teams can now design marketing agents that run on their own data, inside their own guardrails, and coordinate with external AI systems like Claude and ChatGPT through an open Model Context Protocol (MCP) server.

MoEngage, the agentic customer engagement platform used by more than 1,350 consumer brands, launched Merlin AI Custom Agents. Lifecycle marketers and CRM teams can now build their own workflow agents on top of MoEngage data and tools, define the rules each agent runs inside, and watch every step the agent takes. MoEngage is also opening to the enterprise AI stack through an MCP server, so customers can use AI tools like Claude or ChatGPT to access MoEngage data / tools or build external agents leveraging the MoEngage connector.

Most agentic AI in marketing today operates as a black box. The marketer enters a prompt, the system returns an output, and what happened in between stays hidden. For a CMO running hundreds of millions of customer touchpoints across email, push, in-app, and SMS, that opacity is a non-starter. Merlin is built differently. Every Merlin AI agent shows its work. Marketers see the data it pulled, the decisions it made, the channels it touched, and the content it sent. They set the audiences, channels, content rules, and budget limits before the agent runs. Some teams want full autopilot. Others want a hand on the wheel and review every action before it ships. The same agent runs either way.

Marketing Technology News: MarTech Interview with Haley Trost, Group Product Marketing Manager @ Braze

“Marketers have been telling us for two years they don’t want AI that just does the work. They want AI they can see inside, set rules for, and stop when they need to. Merlin is built around that. The reason teams can let our agents run 24/7 and run experiments at a pace they couldn’t touch manually is that the agents only do what the marketer’s guardrails allow. Visibility and control are what make autonomy safe. Teams at SoundCloud, Loblaws, Swiggy, Domino’s, and dozens of other consumer brands shaped what we shipped,” said Raviteja Dodda, Co-founder and CEO of MoEngage.

Most martech AI today is built for one audience: a human marketer typing into a chat box. MoEngage is building for both the marketer and the AI agents already running inside the company. Through the MCP server and agent-callable APIs, MoEngage agents pass context to and take direction from external systems. For enterprises standardizing on a platform, the existing stack stays in place rather than getting replaced.

What’s new

Custom Agents in Merlin AI. Build agents that run continuously on your workflows. Every custom agent runs on MoEngage data, uses MoEngage tools, and operates inside the rules a marketer defines. A full activity log shows every decision it makes. Customer use cases: agents that QA campaigns before they go out, agents that turn a creative brief into a journey ready for review, and agents that produce analytics reports without anyone opening a dashboard.

Marketing Technology News: Cross-Department Collaboration with Marketing Workflow Automation: Enhancing Alignment Between Sales, Customer Service, and Marketing Teams

New Merlin AI agents. Three additions to the agent suite already in Studio:

• In-App Template Generator. Describe the in-app message you want. Merlin returns the App Template with responsive code and interaction logic done.
• Flows Assist. Type the objective of a journey. Merlin returns a multi-stage canvas with triggers, paths, and channel steps wired in. The marketer reviews, edits, and ships.
• Campaign Insights Agent. Ask questions about your campaigns in plain language. The agent surfaces what’s working, what isn’t, and what to change next.

MoEngage MCP Connector. Connect MoEngage to external AI systems like Claude and ChatGPT through the Model Context Protocol. Your own agents read MoEngage context, coordinate with Merlin agents, and take action across your stack without custom integration work.