Application Refactoring
Infosys wanted to move beyond basic AI augmentation in the software development life cycle because developers still had to manually review, validate, integrate, and fix generated code in large enterprise applications, limiting productivity and increasing delivery risk.
Reported impact
Infosys reports materially faster software development and modernization work, including higher productivity in database code generation, API and microservices creation, user interface generation, unit-test creation, database-view migration, and application dependency upgrades, while using iLEAD guardrails to keep generated code aligned with enterprise standards.
Infosys uses GitHub Copilot agent mode as the execution layer for software-development work across enterprise applications and its own IP products and platforms. Architects and developers feed the agent detailed context such as tech-stack information, business rules, user stories, coding standards, prompts, and rule sets so it can plan multistep tasks, generate code, suggest terminal commands, inspect runtime errors, and iterate on its own output through self-reflection. Infosys then wraps this workflow with its iLEAD platform, an AI-first, agent-integrated SDLC environment that supports requirements definition through release management and adds guardrails for maintainability, security, and enterprise standards. Infosys is also developing model context protocol servers inside iLEAD to bridge agents to external systems and expose tools such as databases, design patterns, and industry- or technology-specific endpoints, so generated changes can be better integrated, reviewed in smaller chunks, and handed off across architect, developer, and DevOps roles.
Tech-stack details, business rules, user stories, coding standards, and prompts are provided to GitHub Copilot agent mode as context. The agent works against the existing code base, terminal, and operating system, iterating on generated code and runtime errors. Within iLEAD, model context protocol servers expose databases and design-pattern resources, and reviewed output then moves through SDLC and release-management workflows for architects, developers, and DevOps teams.
According to our analysis, which includes experiments from both senior architects and junior developers at Infosys, agent mode can achieve between 80% and 90% improvement in database code generation; between 60% and 70% improvement in generating application programming interfaces (APIs) and microservices; and up to 60% improvement in generating user interface code.
At Infosys we have adopted GitHub Copilot at scale and its agent mode is particularly exciting.
Creating end-to-end unit test cases took about 45 minutes with AI assistance, compared to a couple of weeks it would have required manually.
Migrating Oracle database views to SQL views was completed in just 10 minutes instead of the approximately three days usually needed.
The platform ensures that the agent-generated code adheres to best practices, incorporates strong security measures, and is straightforward to maintain.
This agent is capable of iterating on its own code, understanding its own errors, and then fixing them itself through self-reflection.
To facilitate this, we are also developing model context protocol servers — software components that act as bridges between AI agents and external systems — within the iLEAD platform, providing tools and resources for agents.
To help in this, Infosys leverages the iLEAD platform. This is an AI-first, agent-integrated platform providing end-to-end SDLC assistance tailored for key personas including architects, developers, and DevOps teams, supporting the journey from requirements definition to release management.
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
Microsoft Sample Oct 17, 2024 H&R Block H&R Block wanted to build its first AI-assisted tax preparation solution for individual filers. Half of 150 million US taxpayers file their own returns. H&R Block, a tax preparation company, focuses on… Azure App Service · Azure Functions · Azure OpenAI Service 
OpenAI Sample Oct 8, 2026 Endava Endava needed to remain relevant in an AI-centric market by redesigning software delivery and cross-functional workflows, not just adding point tools to existing engineering processes. 11,000-person global workforce AI agents · ChatGPT Enterprise · Codex 
ServiceNow Sample Modified · Jul 30, 2026 Syslabo Syslabo needed to replace a 30-year-old legacy sales support stack, remove silos between separate sales tools, and give business users a practical way to build and improve internal applications without long specialist… 2x Faster onboarding for App Engine with AI Agents AI Agents · App Engine · Now Assist for Creator 