ITS Global’s AI Solutions

Legacy System Modernization

Achieve Quickly and Cost-Effectively

CodeMigrateAI

Challenges in Advancing Modernization

End of Maintenance

Unable to adequately support systems whose maintenance has ended or is scheduled to end.

Black Boxization

Years of operation have made the system architecture opaque, making it difficult to understand, modify, or maintain.

Aging Technology

The technologies in use have become outdated, making it difficult to secure engineers and skilled personnel.

Modification Costs

The system architecture is complex, making modifications and feature additions time-consuming and costly.

ITS Global’s Solutions

Delivering modernization through a unique solution that combines ITS Global’s diverse services.

AI-Powered Analysis (CodeMigrateAI)

Utilize generative AI to analyze the structure of existing source code, reconstruct the code, and redefine and update the system specifications.

What Can Be Achieved

・Analyze the specifications of a black-boxed system and visualize and reconstruct them.
・Optimize and redesign code based on To-Be requirements
・Support decision-making and planning for advancing modernization

Service Flow

STEP 1

Analysis of Existing Programs

STEP 2

Design Optimization

STEP 3

Automatic Code Generation

STEP 4

Unit Testing

Screen design by UI designers

Enhance business performance with excellent usability.

Extraction of Common Specifications Using AI

Leverage AI to analyze multiple specifications and requirements, and derive common specifications.

Factory A Specifications

Factory B Specifications

Factory C Specifications

CodeMigrateAI

Common Specifications

Modernization Approach

Step 1

Analysis & Planning

CodeMigrateAI（CMA）

Feasibility Study / Planning

Based on the analysis of a subset of functions, estimate the overall system size and define the required effort and project schedule.

CMA Compatibility Validation

Analyze the current state of the application and assess the feasibility of applying CMA.

CMA Overall Plan

Conduct additional validation and consider optimization of the To-Be architecture, then provide an overall size estimate and a proposed modernization plan.

Agile Development

Step 2

Implementation

Modernization and Migration by CMA (Full-Scale Development)

Visualize a black-boxed system and transform it into a white-box system.
No Changes to Functionality
Existing functionality is maintained while the interface is refreshed with a newly designed UI.

Waterfall Development

Step 3

Production Launch

Equipped with New Features

01 Operations & Maintenance

02 Phase 2

Implementation Benefits

Sales Management System

Project Outcomes

Significantly reduce coding time while streamlining the processes of design documentation and source code generation, thereby enhancing both development speed and quality.

Programming Language

RPA -> JAVA

Source Code Size

Average of approximately 18K steps (about 12 modules)

Preliminary Estimate (Approximately 35% Reduction)

Without Generative AI: Approximately 120 person-months / 150 million JPY

With Generative AI: Approximately 78 person-months / 97 million JPY

Import/Export Management

Project Outcomes

By leveraging generative AI, engineers can handle development and modifications without requiring COBOL expertise. Additionally, standardizing the processes for design document and source code generation helps reduce quality variations.

Programming Language

COBOL (AS/400) -> Java

Source Code Size

Average of approximately 11K steps (about 7 modules)

Preliminary Estimate (Approximately 35% Reduction)

Without Generative AI: Approximately 28 person-months / 35 million JPY

With Generative AI: Approximately 18 person-months / 22 million JPY

Import/Export Management System Overview

Supported Languages & Domains

Supported Languages (Input / Output)

Input (Legacy)

COBOL PL/I VB6 RPG Java (Old) PHP (Cake/Symfony (Old))

Output (Modern)

Java (Latest) TypeScript Python Go PHP (Cake/Symfony (Latest))

Frequently Asked Questions

Test code is generated automatically; however, all deliverables—including code and design documents—are reviewed by humans to ensure quality.

There is no risk of information leakage, as opt-out settings are applied to prevent input data from being used for training.

Since it varies depending on the project’s scope and details, please feel free to contact us for consultation.