Infosys and OpenAI Are Targeting Legacy Modernization With Codex and Agentic AI

Infosys announced a strategic collaboration with OpenAI on April 22, 2026, to help enterprises use OpenAI models and products such as Codex for software development and modernization. The partnership combines OpenAI’s technology with Infosys Topaz Fabric, which Infosys describes as a composable and open agentic services suite.

The focus areas are practical: software engineering, legacy modernization, DevOps automation, ecommerce, and engineering-led domains. Infosys says the collaboration will combine Codex, workflow automation, prebuilt agents, its poly-AI architecture, and enterprise governance to improve engineering productivity and reduce time-to-market.

This is an important India IT services story because Infosys is trying to productize AI-assisted modernization. Indian IT companies have spent decades maintaining and upgrading enterprise systems for banks, insurers, retailers, manufacturers, telecom companies, and public-sector clients. Much of that work involves old codebases, outdated platforms, missing documentation, weak test coverage, and business logic that exists only inside legacy systems.

Codex fits into that problem because modernization is not just about writing new code. Teams first need to understand the existing system, identify dependencies, document hidden logic, generate tests, find risky code paths, refactor modules, detect vulnerabilities, and then migrate carefully. Infosys is positioning Codex as a tool that can sit inside those delivery workflows instead of being used only by individual developers.

OpenAI is also scaling Codex through global systems integrators. OpenAI says it is launching Codex Labs and working with partners including Accenture, Capgemini, CGI, Cognizant, Infosys, PwC, and Tata Consultancy Services to help enterprises identify and deploy Codex use cases across the software development lifecycle.

That systems-integrator route matters. Large enterprises rarely adopt engineering AI through random tool usage alone. They need workflow design, security review, repository access control, testing standards, team training, integration with CI/CD pipelines, and governance. Infosys can package Codex into repeatable delivery playbooks for clients that already use its modernization, cloud, DevOps, and application maintenance services.

Legacy modernization is the clearest use case. Many enterprises still run important workloads on older systems because migration is risky, expensive, and slow. Codex can help engineers understand old code, generate documentation, write test cases, suggest refactors, and support incremental migration. The final responsibility still stays with engineering teams, but AI can reduce the manual effort required to start and review modernization work.

DevOps automation is another practical area. Enterprise software delivery involves build scripts, deployment pipelines, infrastructure configuration, monitoring checks, test automation, release notes, and incident follow-ups. Codex can help teams generate pipeline code, explain deployment failures, suggest fixes, and produce checklists for repeatable release workflows. Infosys’ announcement specifically names DevOps automation as one of the early focus areas.

Vulnerability detection is also relevant. OpenAI’s revenue chief Denise Dresser said Infosys’ software transformation experience can help enterprises deploy Codex across legacy code modernization, code review automation, vulnerability detection, and application development. For clients in finance, healthcare, retail, telecom, and manufacturing, this is valuable because older applications often carry security debt that is difficult to find manually.

The ecommerce use case is also worth noting. Ecommerce systems usually include product catalogues, checkout flows, payment integrations, order management, promotions, CRM integrations, search, personalization, and analytics. Codex and agentic workflows can support feature development, test generation, bug fixing, API documentation, migration work, and automation across these connected systems.

Infosys Topaz Fabric gives the partnership a platform angle. Infosys is not only saying that its teams will use Codex. It is combining Codex with its own agentic services suite, governance model, workflow automation, and enterprise delivery structure. That is how IT services companies are likely to sell AI-assisted modernization: as a managed delivery capability with tools, people, reusable agents, and process controls.

For Indian readers, this points to a larger shift in IT services. The work is moving from manual effort-heavy modernization to AI-assisted delivery systems. Clients will still need engineering partners, but they will increasingly expect faster discovery, faster test creation, faster refactoring, better documentation, and more automation in routine engineering tasks.

This does not remove the need for skilled developers. Enterprise modernization requires architectural judgment, business-domain understanding, security review, data migration planning, production-risk management, and stakeholder coordination. Codex can help generate and reason through work, but human engineers still need to verify, test, approve, and release changes.

The biggest risk is careless adoption. AI-generated code can be wrong, insecure, or unsuitable for a specific business system. Enterprise use needs code review, automated testing, static analysis, dependency checks, audit trails, permission control, and clear human ownership. Infosys’ emphasis on enterprise governance is important because modernization work usually touches systems that companies cannot afford to break.

Infosys and OpenAI’s collaboration shows how AI coding agents are becoming part of enterprise delivery models. For clients, the value will depend on whether Codex can reduce modernization timelines, improve code review, increase test coverage, detect risky patterns, and make old systems easier to change. For Indian IT services firms, this is the next competitive layer: turning AI-assisted engineering into a repeatable, governed modernization offering.