Most organizations don't lack a modernization strategy. They lack working software and measurable results. Meanwhile, legacy applications keep getting more expensive to carry and harder to change. Doing nothing isn't the low-risk option. It's the compounding one. The obstacles go beyond legacy code: data that is locked in old systems and infrastructure that cannot scale for AI hold organizations back too. 

The question is no longer whether to modernize, but what each application should become, and how quickly the organization can get there. 

Why modernization stalls 

An application portfolio is rarely one thing. Mainframe workloads and virtualized estates each call for a different set of skills and tooling. A team that excels at COBOL is not necessarily the team that converts databases, and few organizations have every skill set in-house. Modernization is broad; Extracting business logic from legacy or mainframe applications requires different tooling and skills than upgrading a Java or .NET application than AI-assisted VM migrations. 

At the same time, the pressure has shifted. Business leaders increasingly want these routes moving in parallel rather than one application per budget cycle. Doing that well takes more than a collection of point projects. 

There's a second trap: stopping at the code. An application that compiles and runs in the cloud isn't necessarily modernized. If the surrounding infrastructure, licensing and compute choices are carried over unchanged, performance and cost savings are left on the table. 

What "modernization with optimal hardware" means 

Modernization with optimal hardware treats the application and the hardware it runs on as a single design problem, weaving together WWT's OEM partnerships with our application modernization and cloud expertise. The goal is a solution that is: 

  • Fast. Releases measured in days rather than quarters, with applications that scale on demand.
  • Cost-conscious. Licensing and run costs designed out rather than carried over.
  • Matched to the right compute. Each workload placed on the hardware that suits it, matching application needs to the right compute for the best outcome.

AI is changing the math 

The hardest and most expensive parts of modernization have traditionally been understanding what legacy systems do and proving that the rewritten version behaves the same way. Agentic AI is starting to automate both: reading and documenting large codebases, extracting business rules, assisting with refactoring and generating tests, with engineers steering the work. 

The early numbers are notable. McKinsey estimates AI can make complex modernization 40–50% faster. AWS reports that AWS Transform, its agentic AI service for modernization, has analyzed more than 10 billion lines of code for customers. The direction is clear: re-architecture that was uneconomical two years ago may be viable today. Any modernization business case older than that deserves a fresh look. 

WWT's AI-enabled modernization offerings follow four steps:  

  1. Understand - agents read and document large codebases
  2. Extract - business rules become specs
  3. Transform - automated refactoring, with engineers steering
  4. Verify generated tests prove equivalence.

A repeatable path, not a one-off project 

At World Wide Technology (WWT), we approach modernization in three stages. First, prove the approach on one or two applications while the platform foundation is built once. Then industrialize it, with parallel teams working a backlog in pattern-based groups so each repeat goes faster. Finally, shift ownership to product teams so modernization becomes continuous instead of a one-time event. 

1. Prove 

Lighthouse 

1–2 apps modernized end to end while the platform foundation is built once. 

2. Industrialize 

Factory 

Parallel teams work a backlog in pattern-based groups. Each repeat gets faster. 

3. Operate 

Continuum 

Apps move to product-team ownership. Modernization becomes continuous. 

This approach aligns with our decades of enterprise infrastructure experience, our AWS expertise and partnership, and the Advanced Technology Center (ATC). Before committing, customers can test target architectures, compare compute and hardware options from our OEM partners, and validate performance and cost against their own workloads. That turns modernization decisions from assumptions into evidence, and helps teams make the right technology decisions faster. In the pilot stage, that means proving a migration pattern, a licensing change or a compute choice in the ATC first, so each repeat in the industrialize stage starts from validated results rather than assumptions. 

Where to start 

The most useful first step is rarely a migration. It is a clear view of the portfolio: which applications to retire, rehost, refactor or rebuild, and in what order. A short, focused workshop with application and business owners can align on drivers and candidate applications, and an AI-assisted assessment can turn that into a data-backed disposition for each one. From there, the plans make the first pilot applications concrete, with funding options such as the AWS Migration Acceleration Program identified early. 

Organizations that start this way tend to avoid the two common failure modes: a strategy that never turns into action, and a migration that moves the problem without solving it. 

Continue the conversation at AWS re:Invent 

This only scratches the surface. To go deeper, join Rashid Sajjad, Technical Solution Architect III, and Cameron Hatten, AWS Practice Manager, at AWS re:Invent for a working session on application modernization with optimal hardware. The session, Infrastructure Aware Modernization, takes place on December 2nd from 12:30 – 1:00 at Booth #1445. 

Not attending Re:Invent? Reach out to our Cloud Practice team and AWS experts to start the conversation. Every application you carry over unchanged keeps costing you. Let's decide what each one should become. 

Connect with WWT's Cloud Practice to assess your application portfolio and build a practical roadmap for modernizing on AWS. Start Today

Technologies