DocumentationAEMUser Guide

Development Agent Overview development-agent-overview

Last update: June 17, 2026
  • Applies to:
  • Experience Manager as a Cloud Service

CREATED FOR:

  • User
  • Admin
  • Developer

As part of the Brand Experience Agent, the Development Agent helps traditional AEM Java-stack developers and administrators create, debug, deploy, and optimize code more efficiently.

It supports the following jobs, which are accessible through the AI Assistant’s conversational interface.

  • Cloud Manager Job: read-only operations, including listing of programs and environments, and pipeline status
  • Pipeline Troubleshooting Job: debug failed pipelines
  • Quiet Hours and Update Free Periods Management Job: view, create, and edit Quiet Hours and Update Free Periods
  • Replication Troubleshooting Job (Beta): debug replication-related issues such as blocked queues.
NOTE
Developers will also find these AI-powered features useful:
IMPORTANT
AI-generated responses may be inaccurate or misleading. Be sure you double-check suggested fixes and responses.
See also Adobe Experience Cloud Generative AI User Guidelines.

You can email development agent–specific feedback to aem-devagent@adobe.com.

Cloud Manager Job cloud-manager-job

Find information about your AEM programs and environments, including:

  • listing programs and environments
  • listing environment variables
  • finding the names of pipelines and current execution status and step details
  • retrieving links to logs that can be downloaded

Sample Prompts sample-cm-job-prompts

Prompt
Result
List all my AEM Cloud Service programs
Lists programs that you have access to.
Get details for program 12345
Retrieve details about the program.
List environments in program 12345
Lists environments in the program.
Get logs for the production environment
Retrieves links to various AEM, dispatcher, and CDN log files so they can be downloaded for debugging or other purposes.
list pipelines for program 12345
List pipelines in the program.
What’s the status of the current pipeline execution?
Responds with status of the pipeline.
Get me the build log links for pipeline execution 12345
Retrieves links to pipeline build logs for a specific pipeline execution.

Quiet Hours and Update Free Periods Management Job control-updates-job

View, create, and edit Quiet Hours and Update Free Periods directly through the AEM AI Assistant.

The key benefit is fewer scheduling errors. As you make a request, the assistant guides you through what is possible and flags the limits that apply, such as the three-period cap, the mandatory one-week gap between periods, and the planned maintenance exclusion windows you cannot schedule over.

So instead of discovering a constraint after a failed configuration, Business Owners and Deployment Managers are steered to a valid schedule in the same conversation. This protects critical business windows from automatic maintenance updates while reducing back-and-forth and misconfiguration.

Sample Prompts sample-updates-prompts

Prompt
Result
What’s the current update schedule for program 12345?
Results in a listing of the current AEM update rules.
Block AEM updates from 9 AM to 5 PM EST for program 12345
Sets up a rule so AEM updates aren’t applied during standard work hours.
Remove the daily update block for program 12345
Removes the rules preventing AEM updates.
Pause AEM updates starting in two weeks for program 12345
Creates a rule to prevent AEM updates.
My program keeps getting updated at inconvenient times. What options do I have?
Responds with information about how to set rules to control the AEM update schedule.

Pipeline Troubleshooting Job cloud-manager-pipeline-troubleshooting

This job can retrieve pipeline statuses and help you troubleshoot failing build steps by suggesting fixes, saving time when debugging AEM as a Cloud Service deployments to development, stage, and production environments. It examines build logs and related code to recommend a fix that you can apply manually.

https://video.tv.adobe.com/v/3478006?quality=12&learn=on
NOTE
Pipeline Troubleshooting is limited to Full Stack pipelines (Deployment and Code Quality), and Web Tier Config Pipeline.

Follow along a tutorial to learn how to use the Development Agent to troubleshoot pipeline failures.

Access the Development Agent Through Cloud Manager how-to-access-the-agent

You access the Development Agent through the AI Assistant found in user interfaces including Cloud Manager or Experience Hub.

  1. To get started, click Adobe Experience Cloud to open its home page.

    Adobe Experience Cloud home page

  2. In the left rail, under the Services heading, click Cloud Manager.

    The Experience Hub left rail showing Cloud Manager listed under the Services heading

    note important
    IMPORTANT
    The widgets, tools, and artifacts shown depend on the user persona, entitlements, and AEM deployment type (AEM as a Cloud Service or Managed Services 6.5/6.5 LTS).

  3. In the left rail, under Program, click Overview icon .

  4. On the Program Overview page, in the Pipelines card, click a pipeline.

    Selected pipeline

  5. In the Build and Code Scanning page, note the failed pipeline.

    Pipeline failure as seen in the Build and Code Scanning page

  6. Near the upper-right corner of the AEM user interface (either from Cloud Manager pages or the author instance of the AEM environments), click the AI Assistant icon.

    AI Assistant icon on the toolbar

    See also AI Assistant in AEM.

  7. In the AI Assistant panel text box near the bottom, type your question or prompt, then press Enter or click Send icon .

    For example:
    In the “eda-org-01-no-access” program, analyze the failure to the “no-access” pipeline and troubleshoot.

    The prompt results in the following response.

    AI Assistant prompt and resulting response

Troubleshoot Directly from a Failed Pipeline Execution troubleshoot-with-ai-button

When a pipeline execution fails, Cloud Manager also surfaces a Troubleshoot with AI button directly on the pipeline execution page. This is the fastest way to start a troubleshooting session because the failed execution is automatically passed as context to the AI Assistant — no manual prompt entry is required.

  1. In Cloud Manager, open the failed pipeline execution. The status banner displays Failed and the Troubleshoot with AI button appears near the upper-right corner of the page.

    Failed pipeline execution page showing the Troubleshoot with AI button and the AI Assistant panel with a pre-loaded analysis

  2. Click Troubleshoot with AI.

    The AI Assistant panel opens on the right side of the screen. The assistant automatically references the failed pipeline execution and begins its analysis, identifying the failing step and suggesting a fix you can apply manually.

  3. Review the response and, if needed, continue the conversation in the AI Assistant text box to ask follow-up questions or request more detail.

Troubleshoot from the Experience Home Failed Pipelines Widget troubleshoot-from-experience-home-widget

Experience Home includes a Failed Pipelines widget that gives you an at-a-glance view of pipeline failures across your programs without requiring you to navigate into Cloud Manager first. Each row in the widget represents a failed pipeline and shows the pipeline name, last run date and time, duration, and the step that failed. A Troubleshoot with AI button is available inline for each entry.

NOTE
The Failed Pipelines widget is visible only when the Admin & IT role is selected in Experience Home. If you do not see the widget, verify that your role is set to Admin & IT using the role selector in the upper-right corner of the page.

The Failed Pipelines widget on the Experience Home page, showing a failed pipeline entry with a Troubleshoot with AI button

  1. Open Experience Home, click on Experience Manager and scroll to the Failed Pipelines widget.

  2. Locate the pipeline you want to investigate, then click Troubleshoot with AI in that row.

    The AI Assistant panel opens with the failed pipeline execution pre-loaded as context. The assistant begins its analysis automatically, identifying the root cause and suggesting a fix.

  3. Review the response and, if needed, continue the conversation in the AI Assistant text box to ask follow-up questions or request more detail.

Permissions permissions

The pipeline troubleshooting job requires either the Cloud Manager - Developer role or the Cloud Manager - Program Manager role.

Sample Prompts sample-pipeline-prompts

Prompt
Result
Troubleshoot my failed pipeline
Performs an analysis of why a pipeline failed; if it is unclear which pipeline is being referred to, additional questions will be asked to the user.
List my failed pipelines for program Main Program.
While results may vary, this prompt outputs a table of failed pipelines, with a follow-up suggestion to reference a specific pipeline to analyze.
Analyze my failed pipeline called “Dev Pipeline.”
This prompt results in an analysis of the failed pipeline with suggestions to fix. If there are multiple failures, additional questions will be asked of the user.
Troubleshoot pipeline execution 1234567
By providing an exact pipeline execution id, a pipeline analysis is performed.

Out-of-Scope Features out-of-scope-features

Pipeline troubleshooting operates on the Build & Unit Testing step and Code Scanning step in Full Stack Deployment and Code Quality pipelines. It also supports web tier config pipelines.

For other pipeline types and steps, debug failures by downloading and inspecting the logs. See Access and Download Logs for more information.

Replication Troubleshooting Job (Beta) replication-troubleshooting-job

Debug replication-related issues such as blocked queues.

Please email aem-devagent@adobe.com for access to the beta program.

recommendation-more-help
experience-manager-cloud-service-help-main-toc