Beyond the Black Box: MoEngage Unveils Controllable AI for Marketers

SAN FRANCISCO, CA – June 03, 2026 – Customer engagement platform MoEngage today launched a new suite of tools designed to pull back the curtain on artificial intelligence in marketing, directly challenging the opaque "black box" nature of many current industry solutions. With the release of Merlin AI Custom Agents and an open architecture, the company is betting that the future of marketing automation lies in transparency, control, and interoperability.

For B2C lifecycle and CRM teams at the helm of campaigns reaching millions, the new offering allows them to design bespoke marketing "agents" that operate on their own data, within marketer-defined rules, and with a complete, auditable log of every action taken. This move aims to transform AI from a mysterious black box into a trusted, transparent co-pilot for brands.

Unboxing the 'Black Box' of Marketing AI

The promise of AI in marketing has always been alluring: hyper-personalization at scale, predictive insights, and automated efficiency. Yet, for many Chief Marketing Officers, the reality has been a double-edged sword. As AI systems have grown more powerful, they have also become more opaque, creating a significant trust gap. A recent Gartner survey found that 53% of consumers distrust AI search results, while 75% expect brands to disclose when they are using generative AI. This consumer skepticism mirrors a deep-seated anxiety within marketing departments.

When an AI recommends a specific customer segment or generates a campaign, marketers are often unable to explain why that decision was made. This lack of visibility creates risks related to brand safety, regulatory compliance under rules like GDPR, and the fundamental ability to measure ROI. For a CMO managing hundreds of millions of customer touchpoints, that opacity is, as MoEngage's announcement states, a "non-starter."

MoEngage's Co-founder and CEO, Raviteja Dodda, confirmed that this was the primary driver behind the new platform. "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," Dodda said. "Merlin is built around that."

Building Trust Through Transparency and Control

The core of the Merlin AI Custom Agents is a framework built on visibility. Instead of simply entering a prompt and receiving an output, marketers can now define the sandbox in which their AI agents operate. They set the guardrails for target audiences, communication channels, content rules, and budget limits before an agent is deployed.

Every action the agent takes—from the data it pulls to the decisions it makes and the content it sends—is recorded in a full activity log. This creates an auditable trail that demystifies the AI's process. The system is designed for flexibility, catering to teams that want to run campaigns on full autopilot and those who prefer a more hands-on approach, reviewing every AI-suggested action before it ships. According to Dodda, this is the key to unlocking AI's true potential. "Visibility and control are what make autonomy safe," he noted.

This approach directly addresses the concerns of marketing leaders who, according to industry research, feel they are losing control over core functions as platform algorithms become more autonomous. By putting marketers back in the driver's seat, the platform aims to enable teams to experiment and scale at a pace they couldn't manage manually, but without sacrificing governance.

An Open Ecosystem for the Enterprise AI Stack

Perhaps one of the most significant aspects of the launch is the introduction of an open Model Context Protocol (MCP) server. In a martech landscape often characterized by walled gardens, MoEngage is building for interoperability. The MCP architecture allows customers to connect external AI systems, such as Anthropic's Claude or OpenAI's ChatGPT, directly into the MoEngage platform.

This means an enterprise that has standardized on a specific large language model can build its own external agents that read MoEngage context, coordinate with Merlin agents, and take action across their entire tech stack without complex, custom integration work. For enterprise architects and CIOs, this is a critical feature. It ensures that their existing AI investments are leveraged, not replaced, helping to bridge the "CMO-CIO leadership divide" that analysts at Forrester have identified as a major hurdle in AI adoption. The goal is to allow the existing stack to stay in place, with MoEngage acting as the agentic engagement layer.

From Strategy to Execution: Agentic AI in Action

Beyond the high-level architecture, MoEngage also released three new pre-built agents to demonstrate the practical power of the platform. These tools are designed to translate high-level marketing objectives into tangible outputs with minimal human effort:

• Flows Assist: A marketer can type the objective of a customer journey, and the agent returns a complete, multi-stage canvas with triggers, decision paths, and channel steps already wired together for review.
• In-App Template Generator: Marketers can describe the in-app message they want, and the agent generates the responsive code and interaction logic automatically.
• Campaign Insights Agent: This agent allows marketers to ask questions about campaign performance in plain language, surfacing insights on what's working and what needs to change, without digging through dashboards.

These agents, along with the ability to build custom ones, represent a shift from using AI for simple tasks like copywriting to deploying it for complex, end-to-end workflow automation. They showcase how agentic AI can handle everything from quality assurance on campaigns to producing analytics reports, freeing up marketing teams to focus on strategy.

Navigating the Future of Agentic Marketing

MoEngage's launch comes at a time of immense appetite for agentic AI. Recent Forrester research indicates that a staggering 88% of B2B decision-makers are already adopting or planning to adopt AI agents. While martech giants like Salesforce and Adobe have long integrated their own powerful AI engines—Einstein and Sensei, respectively—MoEngage's explicit focus on transparency and its open MCP architecture offers a clear point of differentiation.

By showcasing the new platform at Salesforce Connections in Chicago this week, the company is signaling its intent to operate within, not just adjacent to, the broader enterprise ecosystem. The move suggests that the next wave of innovation in marketing technology will not just be about what AI can do, but how much we can trust and understand it. For brands navigating the future of customer engagement, this shift towards transparent, controllable, and interconnected AI represents a critical step forward.