Accelerate Region expansion with the AWS Knowledge MCP server

Organizations are accelerating their global cloud footprints to deliver exceptional customer experiences and enhanced application performance worldwide. As teams scale across AWS Regions, there is opportunity to use hundreds of AWS services with consistent functionality across Regions.

The AWS Knowledge Model Context Protocol (MCP) server helps provide programmatic access for deep understanding of AWS capabilities and their availability across AWS Regions. By combining the deep research of AI agents with MCP Servers, you can automate Regional expansion analysis and generate actionable migration strategies in hours instead of weeks.

In this post, we show you how to build a multi-agent AI system using the new Regional availability tools in the AWS Knowledge MCP server to automate Regional expansion analysis and planning. These specialized agents run locally to analyze AWS CloudFormation and AWS CloudTrail logs within your environment. The agents then query AWS service, feature, API, and CloudFormation resource availability data through AWS Knowledge MCP tools to identify Regional capability gaps and generate actionable migration strategies. All customer data remains within your local environment and is not transmitted to external systems. By automating this analysis, teams can focus on implementation rather than spending time on manual research and compatibility validation.

In this post you will learn how to implement AI agents that can:

• Analyze your AWS infrastructure across Regions
• Identify service dependencies and feature requirements
• Perform deep research against target Regions when deploying workloads and functionalities
• Generate detailed migration strategies with implementation guidance
• Maintain continuous visibility into Regional capability changes

Solution overview

Building production-ready AI agents for Regional expansion planning requires a systematic approach that combines infrastructure intelligence using deep research on Regional availability with automated guidance to help decision-making. We demonstrate how to combine the AWS Knowledge MCP server along with the AWS CloudFormation MCP and CloudTrail MCP servers to perform deep research with AI agents to automate Regional expansion planning.

The solution implements a multi-agent orchestration pattern built on the Strands Agents SDK, where specialized agents collaborate to deliver comprehensive Regional expansion insights. Each agent serves a specific purpose in the analysis to discover your AWS infrastructure, analyze usage, check for Regional service availability and generate detailed reports with specific migration strategies and implementation timelines.

MCP tools integration

The multi-agent AI system uses several AWS MCP servers to access comprehensive service and Regional availability data. The following table shows the MCP tools used and their specific functions in the analysis:

Architecture

The following architecture diagram shows how the multi-agent system, orchestrated through the Strands Agents SDK, interacts to analyze your infrastructure and generate expansion recommendations to reduce deployment risks while accelerating time-to-market for global applications.

The workflow consists of five key components:

1. The Orchestrator agent coordinates the analysis workflow using the Strands Agents SDK.
2. Infrastructure Discovery analyzes CloudFormation stacks and CloudTrail logs using AWS CloudFormation MCP and CloudTrail MCP servers.
3. AWS Knowledge MCP provides deep research on service availability and features.
4. AI agents process the data to identify gaps and generate recommendations.
5. The system produces three reports: Executive summary, Technical analysis, and Implementation plan

Prerequisites

Before getting started, please make sure you have:

1. An AWS account with programmatic access configured through the AWS CLI (version 2.0 or later)
2. AWS CloudTrail enabled in your source Region for usage analytics
3. AWS CloudFormation Stacks deployed that represent your infrastructure
4. Access to the Claude 4.0 model in Amazon Bedrock
5. Strands Agents SDK (version 1.0 or later). For more information see the Strands Agent SDK User Guide.
6. Python 3.11 or later environment
7. AWS IAM permissions

{
    "Version": "2012-10-17",
    "Statement": [{
        "Sid": "CloudFormationAnalysis",
        "Effect": "Allow",
        "Action": ["cloudformation:ListStacks", "cloudformation:DescribeStacks",
            "cloudformation:ListStackResources",
            "cloudformation:DescribeStackResources"],
        "Resource": "*"
    }, {
        "Sid": "CloudTrailAnalysis",
        "Effect": "Allow",
        "Action": ["cloudtrail:LookupEvents"],
        "Resource": "*"
    }, {
        "Sid": "BedrockAccess",
        "Effect": "Allow",
        "Action": ["bedrock:InvokeModel", "bedrock:InvokeModelWithResponseStream"],
        "Resource":
            ["arn:aws:bedrock:*:*:inference-profile/us.anthropic.claude-sonnet-4-20250514-v1:0"]
    }]
}

Solution implementation: Building your multi-Region expansion AI agent system

We’ve architected this solution with an Orchestrator agent that coordinates five specialized agents through distinct analysis phases. This pattern ensures modularity, reliability, and extensibility as your expansion requirements evolve:

1. CFN Explorer – Analyzes all CloudFormation stacks and resources in a Region
2. CloudTrail Explorer -Processes CloudTrail logs for API usage analysis with a default 7-days lookback
3. Capabilities Explorer – Uses AWS Knowledge MCP tools for Regional availability intelligence
4. Multi-Region Expansion Planner – Orchestrates comparative analysis across target Regions
5. Planning Report Generator – Creates executive-ready expansion planning reports

Phase 1: Infrastructure discovery

The CFN Explorer agent performs comprehensive infrastructure analysis:

1. Discovers and analyzes all CloudFormation stacks and their resources
2. Identifies service dependencies and resource relationships
3. Catalogs encryption keys, networking, and security configurations
4. Orchestrator agent can pass any source Region and a list of target Regions as parameters

Implementation insight: The orchestrator accepts any source Region and multiple target Regions as parameters, to support flexible analysis scenarios.

The following is an example of a successful connection to AWS Account with discovered CloudFormation stacks:

I'll analyze all CloudFormation stacks in the us-east-1 region using the xxxx AWS
    profile. Let me start by retrieving the list of stacks and then gather detailed
    information for each running stack.

Tool #1: use_aws
╭─ 🚀 List all CloudFormation stacks ─╮
│                                     │
│  Service:        cloudformation     │
│  Operation:      list_stacks        │
│  Region:         us-east-1          │
│  Parameters:     None               │
│                                     │
╰─────────────────────────────────────╯

Phase 2: Usage pattern analysis

The CloudTrail Explorer agent analyzes actual service usage patterns:

• CloudTrail Explorer processes 7 days of API usage logs by default. This can be configured to a different value as needed.
• Identifies actual service usage patterns beyond deployed resources.
• Maps API dependencies and integration patterns.
• Orchestrator agent can pass any number of days as lookback period for API analysis.

Production tip: Extend the lookback period for comprehensive analysis during peak usage periods.

Phase 3: Regional availability assessment

The Capabilities Explorer agent provides real-time Regional intelligence through deep research

• Capabilities Explorer checks service availability across target Regions
• Validates API compatibility and CloudFormation resource support
• Identifies Regional limitations and feature gaps in target Regions

Phase 4: Gap analysis and planning

The Multi-Region Planner agent delivers action insights:

• Performs comprehensive gap analysis
• Generates specific workarounds and recommendations based on missing functionality
• Creates implementation roadmaps with effort estimates
• Suggests CloudFormation template changes to implement workarounds

Phase 5: Report generation

The Planning Report Generator agent produces comprehensive documentation:

• Executive summaries with business impact analysis
• Technical implementation guides with code examples
• Compliance requirements and risks matrices

Deploy the solution on your local machine

To deploy the solution on your local machine, complete the following steps:

Step 1: Clone the sample-multi-agent-region-expansion-planner repository:

git clone https://github.com/aws-samples/sample-multi-agent-region-expansion-planner.git 
cd sample-multi-agent-region-expansion-planner

Step 2: Install the required dependencies:

pip install -r requirements.txt

Step 3: Configure the AWS profiles and target Regions in src/utils/config.py:

 # AWS CLI profile to access account where workloads are deployed.
 PROFILE_NAME = 'your-profile-name' 

 # Source region where workloads are deployed.
 SOURCE_REGION = 'us-east-1' 

 # Target regions where workloads expansion need to be analyzed and planned.
 TARGET_REGIONS = ['eu-central-1', 'ca-central-1', 'ap-southeast-7'] 

Step 4: Configure the AWS profile and Amazon Bedrock model provider account in src/utils/config.py:

 # Account ID to use Bedrock Models and create AI Agents
 INFRA_ACCOUNT = "123456789012" 

 # AWS CLI profile for auth
 BEDROCK_AWS_CLI_PROFILE = "bedrock-profile" 

Step 5: Run multi-Region expansion analysis:

 python3 src/main.py

Generated insights and deliverables

Upon completion, the system generates three production-ready reports:

1. Executive summary: Business-focused insights with risk assessments
2. Technical analysis: Deep dive into service dependencies and gaps
3. Implementation plan: Step-by-step migration guide with code samples

Executive summary

The executive summary is a business-focused report that provides key insights about Regional expansion readiness. It presents a clear view of service availability across target Regions, identifies critical dependencies, and outlines required architectural modifications. The report consolidates complex technical findings into actionable business intelligence that helps inform expansion decisions.

The following is an example excerpt of the output for the executive summary:

# AWS Multi-Region Expansion Planning Report

## Executive Summary

This comprehensive expansion planning report provides detailed guidance for migrating AWS
    workloads from **us-east-1** to three target regions: **eu-central-1**,
    **ca-central-1**, and **ap-southeast-7**. The analysis reveals significant service
    availability challenges requiring extensive architectural modifications and a
    carefully orchestrated phased deployment approach.

Our analysis reveals a compelling business opportunity with 78% technical readiness
    across target regions, representing a $485,000 strategic investment that delivers
    $2.1M in additional annual revenue by Year 3. The expansion addresses critical market
    demands for data sovereignty, GDPR compliance, and reduced latency while establishing
    competitive differentiation through first-mover advantage in multi-region knowledge
    graph deployment. The technical foundation demonstrates strong compatibility with 94%
    API parity across regions, though strategic workarounds are required for specific
    service limitations in eu-south-1.

Technical analysis

The technical analysis provides engineers with detailed implementation guidance for Regional expansion. This report includes comprehensive service compatibility analysis, specific API modifications needed, and architectural recommendations for services with Regional variations.

The following is an example excerpt of the output for the technical analysis:

# AWS Workload Migration Technical Analysis

### Key Findings

Based on comprehensive analysis of current AWS workloads in **us-east-1** and assessment
    of four potential target regions, this document provides detailed migration
    feasibility analysis for:

- **ap-southeast-7** (Asia Pacific - Thailand) - **65% compatibility** 
- **eu-central-1** (Europe - Frankfurt) - **95% compatibility**
- **ca-central-1** (Canada - Central) - **80% compatibility**

### Current Workload Profile

The source environment (us-east-1) contains **35 CloudFormation stacks** utilizing **21
    unique AWS services** with the following critical components:

- **Neptune Global Database** with cross-region replication
- **SageMaker notebook instances** for ML workloads
- **Lambda functions** (15 total) with VPC configurations
- **S3 buckets** (8 total) with cross-region dependencies
- **IAM roles and policies** (25 roles, 12 policies)
- **ElastiCache serverless** configurations
- **CDK bootstrap resources** with region-specific naming

### Migration Recommendations

| Region             | Migration Status       | Recommended Action                                       |
|--------------------|------------------------|----------------------------------------------------------|
| **eu-central-1**   | ✅ **RECOMMENDED**     | Proceed with migration                                   |
| **ca-central-1**   | ⚠️ **CONDITIONAL**     | Good foundation - missing CI/CD and development services |
| **ap-southeast-7** | ❌ **NOT RECOMMENDED** | Significant service gaps - wait for service expansion    |

Implementation plan

The implementation plan delivers a detailed migration roadmap. It provides step-by-step guidance for implementing the regional expansion, with clear phases and validation checkpoints.

The following is an example excerpt of the output for the implementation plan:

# Multi-Region Expansion Planning Report
### Executive Report for KnowledgeForge Application Expansion

Source Region: us-east-1
Target Regions: eu-central-1, ca-central-1, ap-southeast-7
Analysis Date: Sep 3, 2025

### Regional Availability Analysis Tables

Service Entity | Entity Type | Source Region | eu-central-1 | ca-central-1 | ap-southeast-7 | Status      | Notes
---------------|-------------|---------------|--------------|--------------|----------------|-------------|------
dynamodb       | AWS Service | Available     | Available    | Available    | ? Unknown      | Full Parity | All core APIs supported
neptune        | AWS Service | Available     | Available    | Available    | ? Unknown      | Partial     | Global DB limitations in some regions
sagemaker      | AWS Service | Available     | Limited      | Available    | ? Unknown      | API Gaps    | DescribeNotebookInstanceLifecycleConfig missing
cloudwatch     | AWS Service | Available     | Available    | Available    | ? Unknown      | Full Parity | All monitoring features available
lambda         | AWS Service | Available     | Available    | Available    | ? Unknown      | Full Parity | All runtimes supported
s3             | AWS Service | Available     | Available    | Available    | ? Unknown      | Full Parity | Regional buckets required
iam            | AWS         | Available     | Available    | Available    | ? Unknown      | Full Parity | Global

Environment clean up

Maintain operational hygiene by running the cleanup command so that agents do not consume data from previous analyses.

python3 src/main.py --cleanup

Conclusion: Transforming Regional expansion with agent workflows and deep research

In this post, we demonstrated how the combination of the AWS Knowledge MCP server and AI agents can transform Regional expansion planning from a manual process into an automated workflow. With this multi-agent system teams can:

• Accelerate planning cycles with automated infrastructure analysis and deep research
• Make data-driven decisions using comprehensive Regional capability data
• Generate production-ready implementation strategies with specific code examples

For more information about multi-Region architectures, see AWS Capabilities by Region. For more information about building global applications on AWS, see AWS Global Infrastructure.

We’re excited to see how you use the AWS Knowledge MCP server to accelerate your global expansion. To get started, visit the AWS Knowledge MCP Getting Started page.

Alternatively, you can also add the AWS Knowledge MCP server to your AI agentic IDEs and AI assistants such as Kiro, Cursor, Cline, or Amazon Q for command line to perform similar analysis without using the Strands Agents SDK.

About the authors

Veera “Bhargav” Nunna is a Senior Data Engineer and Tech Lead at Amazon Web Services (AWS) pioneering Knowledge Graphs for Large Language Models and enterprise-scale data solutions. With over a decade of experience, he specializes in transforming enterprise AI from concept to production by delivering MVPs that demonstrate clear ROI while solving practical challenges like performance optimization and cost control. Veera is among the first to launch production-level MCP Servers, tools, and agentic solutions that solve critical problems at AWS and for customers globally. When not working on AI-driven solutions, he enjoys aviation and flies his family in small aircrafts, holding a private pilot’s license.

James Liu is a Senior Product Manager – Technical at Amazon Web Services (AWS), where he leads product development for cloud infrastructure and developer tools. With 10+ years of experience in product management, he specializes in driving product strategy, building cross-functional partnerships, and delivering scalable solutions from conception to launch. James focuses on AI-driven infrastructure optimization, dependency management systems that serve enterprise developers at scale and holds multiple patents in AI/ML. Outside of work, James is passionate about mentoring aspiring product managers and exploring the intersection of emerging AI technologies with practical business applications.

Nikhil Jivan is a software engineer at AWS specializing in knowledge graphs and scalable data platforms. With 3+ years of experience in distributed systems and data engineering, he builds solutions that help customers unlock insights from their data. When not coding, you’ll find him in the gym, at the piano, or exploring a new city.

Sucharitha Boinapally is a Data leader at Amazon Web Services (AWS) with 13+ years of experience leading enterprise-scale data solutions. She pioneers AI-driven infrastructure optimization using knowledge graphs and MCP servers to transform complex AWS interdependencies into strategic intelligence, empowering teams to make data-driven decisions about AWS products. Her expertise in agentic AI approaches and advanced data orchestration enables insights that accelerate AWS’s global infrastructure deployment and service innovation. Outside of work, Sucharitha enjoys growing her own food and spending quality time with her family, and friends.