Act. Decide. Introducing The Next Frontier of Agentic Engagement

  • UPDATED: 30 October 2026
  • 6 minread
Act. Decide. Introducing The Next Frontier of Agentic Engagement
Reading Time: 6 minutes

Marketing teams have never had more data, more channels, or more AI tools. Yet most campaigns still start the same way: pick a segment, build a message, schedule it, and hope the average customer in that segment cares.

At NEXT Autumn’26, we announced the pieces that change that. Agents that do the work across your entire marketing stack. A single Global Agent that acts as your marketing teammate. And AI decisioning from Aampe, now fully built into MoEngage, decides what each customer should see. Agents act. Decisioning decides.

Meet the Merlin Agent Harness

An AI model alone isn’t enough to run enterprise marketing. It needs tools to act, context to stay on-brand, access to your data, and rules it cannot break. The Merlin Agent Harness is that foundation: a purpose-built harness for enterprise marketing, with a shared workspace where marketers and agents collaborate.

It brings together:

  • Agent Builder to create custom agents
  • Tools (MCP) that analyze data and create content, campaigns, and flows inside MoEngage
  • Connectors to the rest of your stack, for two-way agent actions
  • Context Hub to codify your brand
  • Multi-agent orchestration so agents hand work to each other
  • Governance and guardrails that define how AI operates in your ecosystem

The shared workspace follows how campaigns really run. Agents analyze data, draft strategy briefs, build audiences, fetch content and offers, design variants, translate, and run QC and previews. The marketer approves at each handoff and stays in control.

A stronger harness: new tools, connectors, Context Hub and creative generation

This release makes the harness considerably more capable.

  • New tools for analytics and flows. Agents can now run behavior, funnel, retention, session, and source analysis and pull dashboard and chart data. They can search flows, read definitions and versions, pull flow analytics, and pause, stop, retire, or archive flows. Additional tools cover product catalogs and recommendations, and drafting and analyzing SMS campaigns (drafts only, so a human stays in the loop).
  • New connectors. Agents can now work across your tech stack, not just inside MoEngage. Think Google Sheets for analysis, Google Docs for campaign briefs, calendars and events kept in sync, and connections to Slack, Meta, CRMs, and more. A marketer can ask an agent to analyze a purchase funnel from a sheet, read a brief from a doc, refresh ad creatives, or pull users from a data extension, all in one workflow.
  • Context hub. Codify your brand value proposition, voice, style, and tone once, and make it available to every agent. Instead of re-briefing an AI each time, agents build assets that already sound like you.
  • Creative generation. Easily build on-brand visuals at scale using the context hub. Build visuals for different channels, like email, notifications, website banners, and social media, in different resolutions. 

Custom Agents: one of our fastest-adopted products

Custom Agents let marketers build their own AI teammates, and the response has been remarkable. Teams across industries are already running them in production:

  • A leading stock trading app built a flow count and waterfall generator
  • A national rewards and loyalty program runs a pre-churn touchpoint audit agent
  • An international airline uses an onsite campaign architect and QA coach
  • A Middle East fashion e-commerce brand built an Arabic marketing brief and compliance auditor
  • A global hospital network drafts newsletters from a clinical news scout
  • A major telecom carrier automated dynamic Jinja date scripting

Most agents so far do the analytical heavy lifting. Analytics and reporting account for about 35% of agents built, followed by campaign creation (about 12%), QC, audit and compliance (about 10.5%), and campaign patching and refresh (about 9.5%). Funnel, retention, and churn analysis; flow and journey analysis; price and market monitoring alerts; and segmentation and content block creation round out the list.

Merlin Global Agent: the next evolution

Custom agents are powerful, but marketers shouldn’t have to decide which agent to call or stitch their work together. Merlin Global Agent is the next evolution: one universal agent for B2C marketing, in a unified agentic workspace.

  • One orchestrator. It works across MoEngage and your whole stack: warehouse, CRM, analytics, and more.
  • Proactive and always on. Give it a goal like growing repeat purchases. It flags a metric drop, offers to diagnose it, and suggests the next campaign.
  • A real teammate. Hand it the task, whether that is analyzing performance, creating a campaign, building a segment, or preparing a report.

Global Agent runs on the Merlin Harness, so it inherits your Context Hub, connectors, data, and guardrails from day one.

AI decisioning by Aampe, now within the MoEngage dashboard

Earlier this year, we acquired Aampe, a leader in AI decisioning. Today, we are announcing that it is fully integrated into MoEngage.

The problem: marketing to “Averages”. Brands rely on segments, groups of people with something in common. To write one message for a segment, you end up designing for a persona that blends everyone’s preferences. We call this person the Average: a customer who doesn’t exist. The Average likes a bit of discount, a bit of convenience, a bit of urgency. Nobody in your audience actually behaves that way, so the message lands softly with everyone.

Three things keep brands stuck with the Average:

  1. Thinking in segments, not individuals
  2. Learnings that reset with every campaign
  3. Calendars and schedules deciding the next action, rather than the individual’s preferences

 

The shift: relevance for each individual. Relevance rises when what you offer matches what a customer wants. AI decisioning is powered by two kinds of context:

  • Business context. What matters to you: the goal to optimize for, the content strategy (value proposition, incentive, product, tone), the signals that count as positive (order placed, plan subscribed) or negative (app uninstalled, opt-outs), and what the AI must not do, like frequency limits.
  • Customer context. What each customer wants: their preferences, which messages they respond to or skip, and their preferred channels and timing.

How Aampe learns. It learns per user, so every interaction adds to its understanding of that individual instead of resetting. It learns semantically, understanding what a message is about (discount, convenience, urgency) rather than just which campaign it was. It also measures causally, crediting outcomes to the messages that drove them rather than what would have happened anyway.

What changes for marketers. The workflow moves from segment, build, schedule to strategize and assign. You set the strategy and guardrails. The AI decides content, channel, and timing for each person.

Early use cases show the range. One use case eliminates calendar conflicts by coordinating timing across teams so multiple units never overlap, with a 20% increase in transacting users. Another scales multi-product journeys by mapping 100,000 message variations to each user’s real-time financial stage.

The foundation: data, channels and guardrails

Agents need to see, reach, and be trusted. NEXT also brings upgrades across all three.

Data: what your agents can see.

  • Warehouse Segments get a native query builder. Building audiences on warehouse data used to mean writing SQL. Now marketers can build them with a visual UI, on a zero-copy architecture.
  • Real-time Delta Data sharing syncs only what changed, such as a customer’s lifetime value moving from $500 to $600, and feeds purchase behavior back to the warehouse.
  • Portfolio brings data from all of a brand’s apps and sites into one workspace, so you can engage cross-brand audiences across channels.
  • New integrations let you ingest data from more partners and sources and sync audiences to them.

 

Channels: where your agents can reach.

  • Send PDFs and add deep links inside RCS messages
  • Enrich transactional messages with RCS
  • Dynamic Sender Routing for reliable SMS delivery
  • Trigger in-app messages from a push click

Guardrails: safe and compliant, at scale.

  • Consent-aware email open tracking
  • Smart Bot Management to track bot activity
  • Robust SMS opt-in management
  • Frequency caps with Limit Messages

Conclusion and next steps

If you’re an existing MoEngage customer, contact your account manager to get started with these new capabilities. Schedule a demo here if you’re new to MoEngage!