- Migration and Modernization›
- AWS Transform›
- Custom
AWS Transform custom
Transform any code pattern and crush tech debt with AI-powered custom modernization agent
AI-powered custom transformations for code, APIs, frameworks, and more
AWS Transform accelerates organization-wide modernization through agentic AI that automates custom code transformations. It provides out-of-the-box transformations for common scenarios like Java, Node.js and Python upgrades. The agent also performs custom, organization-specific transformations like version upgrades, runtime migrations, or complex language translations and architectural changes. Through continual learning from code samples, documentation, and developer feedback, the agent delivers high-quality, repeatable transformations without requiring specialized automation expertise, enabling organizations to scale their modernization initiatives effectively.
AWS Transform custom offers both CLI and web experiences. The CLI enables users to define transformations through natural language and execute them on local codebases, either interactively or autonomously. The web experience manages large-scale transformation campaigns, tracking progress across multiple repositories.
Benefits
Tackle the massive technical debt crisis by automating diverse transformations that traditionally consume 20-30% of enterprise software development resources. Drive large-scale modernization initiatives across your organization, executing consistent, repeatable, and high-quality transformations at scale across hundreds of applications. Achieve faster execution with pre-built transformations for common patterns and custom transformations for organization-specific requirements.
The agent learns your specific transformations through documentation, natural language chat, and code samples. The agent comes with 6 out-of-the-box transformation paths, such as Java, Node.js, Python, and AWS SDK updates. Whether you need version upgrades, runtime and API migrations, framework transitions, language translations, or even architecture decompositions, the agent learns, adapts, and executes your unique transformation requirements.
The agent automatically captures feedback and improves over time. The agent learns from every execution, developer feedback, and code change to continually enhance transformation accuracy and effectiveness, ensuring each subsequent transformation becomes more reliable and efficient.
Capture and amplify your organization's transformation knowledge by defining transformations once and using the agent to execute repeatable tasks across your entire organization. Share transformation expertise effortlessly across teams and projects while the agent automatically improves with every execution, reducing knowledge silos and inconsistent implementations. This enables organizations to scale their best practices and institutional knowledge across hundreds of applications, ensuring consistent quality and approach regardless of team or project scope.
Out-of-the-box transformations
The agent includes pre-built transformations for common upgrade scenarios, including:
- Java runtime version upgrades
- Python runtime version upgrades
- Node.js runtime version upgrades
- Java AWS SDK v1 to v2 version upgrades
- Python AWS SDK version upgrades (Boto2 → Boto3)
- Node.js AWS SDK v2 to v3 version upgrades
Impact by the numbers
Up to 85% efficacy rate for out-of-the-box (OOB) transformations such as Java and Node.js version upgrades.
out-of-the-box transformations, including Java, Node.js, Python, and more.
End-to-end automation
Define
AWS Transform offers out-of-the-box transformation definitions for common use cases, such as Java, Node.js and Python upgrades. It also allows you to create custom transformations. To create a custom transformation, a developer familiar with the specific task converses with the agent through language chat and provides the agent with reference materials like documentation and code samples.
Execute
Transformations can be executed interactively, with a human developer overseeing the agent’s work, or autonomously. Autonomous transformations can be triggered with a simple one-line CLI command and can be scripted or embedded into any existing pipeline or workflow. AWS Transform also offers transformation management web experience for monitoring campaigns at scale.
Verify
After a transformation has been executed, user-defined validation steps can be performed. These can be human code reviews, automated validation scripts, or test deployments. If transformed code fails validation, it can be sent back to the agent to fix or flagged for further review.
Learn and improve
The custom agent enables transformation improvement through continual learning. Every time a transformation is completed, the agent automatically identifies knowledge items that it believes will improve results for your transformation tasks later. These can be derived from debugging steps, human input, or general code observations. Transformation owners can review knowledge items that the agent discovers and enable them for future executions.
Customers
Air Canada
"We were struggling with high technical debt for a lot of our codebases, including thousands of Lambda functions running on end-of-life runtimes. We needed to upgrade from Node.js 16 to 20 runtime among other modernization efforts. In a few days, our platform team was able to deploy AWS Transform to coordinate and execute the modernization of all of them, achieving a 90% efficacy rate and an 80% reduction in expected time and costs for the project. We are now making AWS Transform part of our internal standard going forward. AWS Transform is fantastic."
Ray Galipeau, Senior Director, Cloud, Network & Platform Services, Air Canada
Twitch
“We built a transformation on AWS Transform to handle our AWS SDK V1 to V2 Golang migration, and it is giving us an average of 70% acceleration on each application migration. Across 913 repositories, we project saving approximately 2,876 developer days, equivalent to 11 developer years.”
The Twitch Team
QAD
“QAD customers struggled with modernizing from older versions with undocumented Progress ABL customizations to our QAD Adaptive ERP platform. We adopted AWS Transform for our modernization workflow, and the results have been transformative. What used to be a two-week project can now be completed in just three days, enabling 60–70% productivity gains. On average, we're saving 96 developer hours per project and an estimated 7,500+ developer hours annually. Our team has analyzed over 180,000 lines of legacy code with remarkable accuracy. The agent’s continual learning capability improves our transformation quality over time, enabling QAD customers to quickly modernize to the latest QAD Adaptive ERP Platform.”
Sanjay Brahmawar, Chief Executive Officer, QAD
MongoDB
“AWS Transform automates repetitive, error-prone, and repeatable transformation tasks, which in turn will reduce migration complexity and effort while ensuring comprehensive functional and compatibility validation across an entire application stack. This is something that many organizations can benefit from, including MongoDB. Based on some initial insights we have gained, we think AWS Transform has the potential to make a big impact in modernizing and migrating Java applications.”
Melissa Plunkett, Vice President of Product Management, MongoDB
Classmethod
“AWS Transform delivered proven results through two powerful use cases. In the first case, we generated comprehensive documentation for a ColdFusion system in just 30 minutes, a task that would have traditionally required several person-months to understand the specifications. In the second case, we completed a Vue.js 2 to 3 upgrade in less than a day, including several hours of manual work, a task that was originally estimated at one person-month. These results demonstrate that AWS Transform significantly reduces both effort and risk in the initial phases of modernization projects.”
Satoshi Yokota, CEO, Classmethod Inc.
The Gnar Company
“At The Gnar Company, we specialize in large-scale tech debt remediation and modernization, and AWS Transform has been instrumental to our project success, delivering dramatic efficiency gains across multiple client engagements. Recent examples include a customer internal API migration that reduced the timeline by over 60% and an Angular to React migration across several projects that delivered a 75% timeline reduction. We look forward to continued collaboration with AWS as we continue to deliver impactful results for our client partners.”
Mike Stone, Co-Founder, The Gnar Company
Coupang
"At Coupang, we faced the daunting challenge of upgrading 700+ applications to a newer Java version to improve security, boost performance, and enable Graviton adoption. For our initial phase, we selected 70+ applications as our first batch. This type of modernization effort would have traditionally required significant manual effort. However, by leveraging AWS Transform and applying customizations including domain specific configurations, deployment config, JDK settings and fixing missing dependencies, we achieved remarkable results. We successfully transformed all 70+ Java applications in just 2 months with a small team of 5 developers - representing approximately 90% reduction in project timeline compared to traditional manual approaches. AWS Transform have been a game-changer for Coupang, enabling us to rapidly enhance our application ecosystem at scale and stay ahead in the competitive e-commerce landscape."
Ning Zhang, VP of Infrastructure, Coupang
PwC Australia
"We tested AWS Transform in private preview and found it exceptionally developer-friendly with intuitive CLI support and excellent for refactoring assets in development and testing.
We experimented with Python-to-JavaScript conversion for serverless scripting and Playwright-to-Cucumber for test verification - closely mirroring medium-sized delivery projects that transform application languages and tooling to align with enterprise guidelines.
Traditional transformation involves comprehending existing code, selecting target modules, creating repositories, and integrating with cloud solutions like ECS, EKS, and Lambda. Typically, transforming 10,000 lines of code requires 50-80 developer days, with total lifecycle effort reaching 150-180 person days including verification, defect fixing, and overhead.
Our experience indicates AWS Transform can reduce entire lifecycle effort by 25-35% when properly utilized. This solution significantly improves application modernization efficiency across SDLC phases, enabling consulting partners like us to pivot quickly and deliver business value faster and more consistently."
Dr. Binqi Zhang, Managing Director, Engineering, Advisory - Digital, PwC Australia
Deloitte
“Deloitte is leveraging AWS Transform to drive transform-at-scale modernization templates across multiple stacks, including .NET, Java and Node.js. By embedding AWS Transform within a Modernization Factory, we can now apply the four pillars—Knowledge, Transform, Template, and Execute—to enable pattern-based modernization, consistent quality controls, and continuous learning across projects. Our experience of AWS Transform has shown that this process is geared to deliver up to 60% faster modernization throughput, 40% reduction in manual rework. With AWS Transform, Deloitte can now unlock hundreds of engineering hours for innovation and strategic value creation, by automating repetitive transformation tasks and scaling learnings across portfolios”
Jason Howard, Director, Software Engineering, Deloitte
