Skip to main content
PrerequisitesBefore you begin, ensure:
  • Your workspace has access to the Global Agent. For more information, refer to Global Agent: Overview.
  • You have access to the MoEngage data your request depends on, such as campaigns, segments, or analytics.

​
Start a Task

1

Open the composer

Hover over Merlin AI in the left menu and click Home, or click New task to open the composer directly.
2

Select the agent

Click Select agent and choose Global agent. Every task runs on one agent, so select the agent before the task starts.
3

Select a mode

Choose Fast, Balanced, or Expert. Balanced is the default mode, and the mode locks once the task starts, so to use a different mode, start a new task.The Global Agent sets the mode for each task. A Custom Agent sets the mode on the agent instead, so the control does not appear when you select a Custom Agent before the task starts. After the task starts, the control shows the agent’s mode and stays locked.
4

Describe the work

Write what you want in plain language, and then send the message.
The task composer with the Select agent control and the Balanced mode control highlighted below the message box
Each task is a separate conversation. The agent keeps every instruction, tool result, and generated file from the task. You build on your request turn by turn, and the agent carries the earlier detail forward. State the outcome you want, the time window, and the channels or campaigns involved. The agent asks you for any missing detail.
The first message starts the run: the agent begins work as soon as the message arrives.

​
Parts of a Task View

An open Global Agent task with eight numbered callouts marking the breadcrumb, the Visible to workspace switch, the Output files icon, the task menu, the Finished working line, the notification banner, the composer, and the My tasks and Workspace tasks tabs
  1. Agent and task name. The breadcrumb names the agent the task runs on, followed by the task name. To change the name, refer to Rename a Task.
  2. Visible to workspace. Switches whether your colleagues find the task under Workspace tasks. For more information, refer to Change Task Visibility.
  3. Output files. Opens the panel that lists every file the run produced. For more information, refer to Review Task Outputs.
  4. Task menu. Pins, renames, copies a link to, and archives the task. For more information, refer to Task Menu.
  5. Finished working. Expands the steps the agent took, including any error the agent encountered.
  6. Notification banner. Turns on a browser notification for when the task finishes. For more information, refer to Notifications.
  7. Composer. Sends your next message in the same task. The agent picker and the mode control sit below the message box.
  8. My tasks and Workspace tasks. Switches the Recent tasks list between your own tasks and the tasks shared with your workspace. For more information, refer to Recent Tasks.

​
Example: Improving Day 1 Retention

The following walkthrough shows how a task develops across four turns. The values are illustrative.
A Global Agent task named User retention analysis, last 15 days, showing the retention table with Day 1 at 6.5 percent, the diagnosis beneath it, and the target options the agent offers
Each turn builds on the previous turn, so you refine the question rather than restating the background. Publishing the draft stays a step you complete in the campaign builder.

​
Attach Files to a Task

Select the plus icon in the composer to attach files to any message, including the first message. The agent reads each attachment as part of the same turn. Send a CSV of campaign IDs or a brand guideline document, and the agent works from the attached file.

​
Monitor a Running Task

Tasks run in the background. The run continues while you work elsewhere in the dashboard, and you return to the task at any time. The left pane and the All tasks page mark each task with its state:
  • A running task shows a pulsing dot.
  • An unread task shows an orange dot.
  • The orange dot clears when you open the task, and returns when the task updates.
  • A task that is both running and unread shows the pulsing dot.

​
Notifications

A long task often finishes while you are working elsewhere. The hub raises a browser notification when the agent stops.
1

Open Notification settings

Select the notification icon beside Merlin AI Agents.
2

Ask your browser for permission

Select System notification when a task finishes while I’m away, and click Save. Your browser then asks for permission.
A task counts as finished when the agent stops working on the task, whether the agent produced a result, failed, or stopped to ask you a question. The notification appears only while the dashboard is open in a tab and you are working in another tab or window.
Permission belongs to your browser rather than to MoEngage. Once you grant permission, the option stays selected, and you turn notifications off by blocking notifications for the site in your browser settings.
This setting covers browser notifications only. Emails, such as the email a scheduled run sends to the agent’s creator, arrive independently of this setting.

​
Control a Running Task

While the agent works, select Stop response to end the run. A message you send in the meantime waits in a queue that you pause, add to, delete from, or resume. Only the task owner stops a run.

​
All Tasks Columns

For the full history, click View all next to Recent tasks. The All tasks page lists every task with the following columns: When a scheduled run finishes or fails, MoEngage emails the agent’s creator the run’s outcome, Task ID, run date, duration, and next run time. The email leaves out the run’s output. The creator turns the email off with the Email toggle on the agent. For more information, refer to Custom Agents: Overview. To filter, pin, rename, or archive the tasks in this list, refer to Manage and Share Global Agent Tasks.

​
Resume a Task

To continue a task you started earlier:
  1. Open the task from the task list.
  2. Send another message.
The agent keeps the earlier conversation, tool results, and generated files, and continues from where the previous turn ended. You resume the manual tasks that you start. A colleague’s task is view only, and so are your own scheduled and test runs. A notice replaces the message box on those tasks, and Start a new task carries the work forward. For what you do on a colleague’s task, refer to Open a Task. A task stays resumable for 30 days after its last activity. After 30 days, the task stays readable, and you start a new task to continue the work. An archived task opens once you restore the task, and a task leaves the task lists once its Custom Agent is deleted.

​
Review Task Outputs

The task collects every file the agent produces during a run. Select Output files in the task header to open the panel, which lists each file with its format and size. Download a file from the panel, or from the card in the transcript where the agent produced the file. Select a file to review the contents before you save a copy. How a file opens depends on the file format:
  • Images and PDFs: PNG, JPG, GIF, WebP, and PDF files open in the preview pane. A download button sits beside the preview.
  • Text files: Markdown, plain text, log, and JSON files also open in the preview pane. A file larger than 2 MB downloads instead.
  • Office and spreadsheet files: .docx, .xlsx, .pptx, and CSV files download directly. Open each file in the matching application to review the contents.
A marker on Output files shows when new files arrive. From the preview, select Download, or Open in new tab for PDFs, images, markdown, text, and JSON files. The agent writes .docx by default, and produces a PDF when you ask for a PDF.

​
Global Agent Memory

The Global Agent reads and writes two memory stores, and your workspace shares both:
  • Workspace memory: Facts stored once for your workspace. Your Custom Agents use this store as well.
  • Global Agent memory: Facts the Global Agent keeps for its own use, shared by every user in your workspace.
Any task can write to both stores, including a task you keep private. You cannot view, edit, or clear either store.
Memory is shared across your workspace. Anything the agent records in one task is readable by colleagues in their own tasks, so treat memory as workspace knowledge rather than a private note.
Memory holds what stays true between tasks. The agent looks up live figures, such as campaign status, audience counts, and performance numbers, each time you ask. Verify any figure that drives a decision.

​
Permissions and Safeguards

Each task has one owner. The owner sends every message in the task, and colleagues read a shared task without continuing the conversation. For what the agent writes to your workspace, the steps that stay with you, and the workspace a task reads from, refer to Permissions and Safeguards.

​
Next Steps

Manage and Share Tasks

Filter the task list, pin and rename a task, control who sees a task, and archive the tasks you no longer need.

Review What the Agent Covers

See the categories of work the Global Agent handles, from audience sizing to pre-send checks.

Build a Custom Agent

Turn a job you repeat into an agent with its own instructions, tools, and schedule.