Migrating legacy Adobe Experience Manager (AEM) implementations to AEM as a Cloud Service (AEMaaCS) can feel overwhelming—especially when dealing with deprecated APIs, OSGi configuration changes, dispatcher modernization, and cloud-native architectural requirements.

Adobe recently introduced an AI-Assisted Code Migration solution that helps developers automate large portions of this migration process using AI-powered agent workflows, migration skills, and Model Context Protocol (MCP) integrations.

aem-cloud-migration-mcp-ai-assisted-code-migration

In this guide, we’ll walk through the complete process of using Adobe’s AI-assisted migration tooling to modernize your AEM codebase and prepare it for AEM as a Cloud Service.

Understanding Adobe’s AI-Assisted Migration Solution

Adobe’s migration framework consists of two primary components:

1. Cloud Migration Skill

The Migration Skill acts as an intelligent migration assistant inside AI-enabled IDEs. It sources Best Practices Analyzer (BPA) findings, identifies affected files in your project, and applies code transformations pattern by pattern. Basically it can:

Analyze migration findings
Locate affected source files
Apply transformation rules
Refactor code patterns
Process migrations in manageable batches
Pause for developer review before proceeding

The migration process follows a structured workflow where developers select a migration pattern, provide findings, review generated changes, and continue iteratively until all issues are resolved.

Currently Supported Patterns in Migration Skill

Scheduler Migration – Legacy schedulers often rely on APIs that are incompatible with cloud-native environments. The migration skill converts Scheduler injections, Manual scheduler registrations and Legacy scheduling configurations into modern OSGi component-based implementations suitable for AEMaaCS.

Resource Change Listeners – AEM Cloud Service introduces restrictions and behaviour changes around resource observation. The migration skill identifies listener implementations and updates them according to cloud-compatible practices.

Please refer below table for all currently supported Migration Skill Pattern:-

Pattern	What it fixes
`scheduler`	`sling.commons.scheduler`-based jobs incompatible with AEMaaCS’s stateless runtime
`resourceChangeListener`	`ResourceChangeListener` implementations requiring Cloud Service updates
`replication`	Legacy `Replicator` API calls replaced by Content Distribution equivalents
`eventListener`	OSGi `EventListener` implementations updated for AEMaaCS event semantics
`eventHandler`	Synchronous OSGi `EventHandler` services adapted for Cloud Service
`assetApi`	Deprecated `AssetManager` and DAM API calls replaced with supported equivalents
`htlLint`	`data-sly-test` redundant constant comparison warnings in HTL templates
`OSGi configs`	`.cfg.json` conversion, runmode scoping, and Cloud Manager secrets/environment variable extraction

2. Cloud Migration MCP

The Cloud Migration MCP (Model Context Protocol) server connects the AI assistant directly to Adobe Cloud Acceleration Manager (CAM).

Instead of exporting BPA reports manually, developers can retrieve migration findings directly from CAM projects through MCP integration. This enables a more streamlined and up-to-date migration workflow.

MCP Server Endpoint

https://mcp.adobeaemcloud.com/adobe/mcp/cloud-migration

The MCP exposes tools that allow migration findings to be retrieved by:

Migration pattern
Severity level
Project context

This eliminates the need to repeatedly download BPA reports during migration cycles.

Prerequisites Before Starting

Before beginning your migration, ensure you have:

Development Requirements

AEM project opened in your IDE
Maven or Gradle-based AEM project
AI-enabled IDE supporting MCP

Note:- If you are not having AI-enabled IDE. Please visit my previous tutorial for Integrating AEM MCP in your IDE.

Migration Assessment Requirements

One of the following:

Best Practices Analyzer (BPA) CSV export
Cloud Acceleration Manager project with BPA reports uploaded

Access Requirements

Adobe ID with CAM access permissions
Cloud Manager access
Source code repository access

Adobe recommends using CAM with MCP because it provides the most current BPA findings without requiring CSV exports. Assuming that all the prerequisite are confirmed and you have BPA report handy or uploaded to CAM.

Let’s proceed to next section for configuring our cloud migration MCP.

Configure Cloud Migration MCP

To enable direct integration with CAM:

Open your IDE MCP configuration.
Add the Cloud Migration MCP server URL. https://mcp.adobeaemcloud.com/adobe/mcp/cloud-migration
Save the configuration.
Authenticate using your Adobe ID.
Verify that migration tools are available.

Once connected, the migration skill can fetch BPA findings automatically from your Cloud Acceleration Manager projects. Large migration projects often generate hundreds or thousands of BPA findings.

The MCP integration allows teams to prioritize work based on severity levels:

CRITICAL
MAJOR
ADVISORY
INFO

This helps development teams focus on blockers before addressing lower-priority issues.

Note:- One important point to note is that at a time, you can connect/authenticate only to one Cloud manager instance. Like if you want to connect to cloud-migration-mcp then you need to remove the config for mcp-aem that we did in previous post.

Installing Adobe Skills

In order to use migration skill for supported patterns like Schedulers, replication agents as mention in above table. We need to install the aem-cloud-service skill package from the Adobe Skills repository. Follow below steps to install the same in your IDE.

Open the integrated terminal in IntelliJ IDEA and navigate to your AEM project root. Run the following command:-

 npx skills add https://github.com/adobe/skills/tree/main/skills/aem/cloud-service --all

It’s always good to go for full skill package installation, but if you want to go with specific skill installation then during installation, you’ll be prompted to select the skills you want to install.For migration projects, Adobe recommends including:

ensure-agents-md
create-component
dispatcher
workflow

After setup, you can ask your AI assistant:-

Migration Scenario	Example Prompt	Purpose
BPA CSV	`Use the migration skill: scheduler only, BPA CSV at ./reports/bpa.csv`	Uses a Best Practices Analyzer (BPA) CSV report as the source of migration findings and applies a specific migration pattern (for example, Scheduler migration) to modernize code for AEM as a Cloud Service.
CAM via MCP	`Fix replication findings from project <projectname>/<projectId>.`	Retrieves migration findings directly from Cloud Acceleration Manager (CAM) through the Cloud Migration MCP server and automatically identifies, analyzes, and fixes cloud migration issues.
Manual (No BPA)	`Migrate event listener in core/src/main/java/com/example/Listener.java`	Migrates a specific source file when BPA findings are unavailable or when developers want to focus on a particular implementation manually.
OSGi Configuration Migration	`Scan my config files and create Cloud Manager environment secrets or variables.`	Reviews OSGi configurations and recommends converting sensitive values into Cloud Manager environment variables or secrets following AEM as a Cloud Service best practices.
HTL Lint Remediation	`Fix htlLint in ui.apps - scan for data-sly-test redundant constant warnings.`	Scans HTL components for linting and code quality issues, then suggests or applies fixes to align with HTL best practices and cloud-ready code standards.

These skills provide guidance for common AEM Cloud Service development and modernization tasks.

Commit Changes Incrementally

Adobe recommends processing migrations:

One pattern at a time
One batch at a time
One commit per migration pattern

Benefits include:

Easier code reviews
Cleaner Git history
Faster rollback capability
Better testing isolation

This incremental approach is especially valuable for large enterprise repositories.

Benefits of AI-Assisted Migration

Organizations adopting AI-assisted migration can achieve:

Faster code modernization
Reduced manual refactoring effort
Improved migration accuracy
Better alignment with Adobe best practices
Easier prioritization of migration tasks
Enhanced developer productivity

The combination of BPA analysis, Cloud Acceleration Manager integration, and AI-guided code transformations creates a more predictable and manageable migration experience.

Migrating to AEM as a Cloud Service is no longer just a manual refactoring exercise. Adobe’s AI-assisted migration framework introduces an intelligent workflow that combines migration expertise, automation, and cloud-native best practices.

As organizations continue moving toward cloud-native digital experience platforms, AI-assisted migration will play a critical role in reducing complexity, improving efficiency, and ensuring successful AEM Cloud Service adoption.

What’s Your AEM Migration Experience?

Have you started migrating your applications to AEM as a Cloud Service using Cloud Migration MCP and Adobe Agent Skills?

Which migration patterns have you found most challenging—Schedulers, Event Listeners, Replication APIs, or OSGi Configurations?
Have AI-assisted migration tools helped reduce manual refactoring efforts in your projects?
What additional migration skills or MCP capabilities would you like Adobe to provide?

Share your experiences, challenges, and tips in the comments below. Your insights could help other AEM developers and architects accelerate their cloud migration journey.

Spread the love