OpenAI agents reshape Endava, raising stakes for Indian IT
Endava is rebuilding its entire software delivery model around OpenAI's agents and Codex. For India's $200bn IT export machine, the near-shore rival's move is a warning shot.
The News
Endava, the global software engineering and technology services group, has detailed how it is rebuilding the way it delivers client projects around OpenAI's agentic tools. In an account published by OpenAI, the company said it is rolling out ChatGPT Enterprise across its workforce and adopting Codex, OpenAI's code-generation model, to automate parts of the development lifecycle and build what it describes as an "AI-native" culture.
The move is less about bolting a chatbot onto existing teams and more about redesigning the delivery model itself. Endava said it is weaving AI agents into how its people plan, write, test and ship software, with the stated aim of accelerating delivery and automating routine work. For a firm whose business is selling engineering capacity to enterprise clients, that is a notable statement of intent: agents as the organising principle, not a bolt-on.
Why It Matters
The IT services industry is facing its sharpest structural test in more than a decade. The sector grew through two big waves: the offshoring boom of the 2000s, which moved delivery to lower-cost geographies, and the cloud migration of the 2010s, which rewired where software runs. Agentic coding is shaping up to be the third, and it strikes at the economics directly.
The traditional model ties revenue to headcount: more engineers booked, more hours billed. If agents can plan, generate and test meaningful chunks of code, that link starts to loosen. When a delivery house publicly reorganises around agents rather than simply licensing a coding assistant, it signals that the billing and staffing playbook is up for renegotiation.
There is precedent for how fast this can move. When GitHub Copilot arrived in 2022, it reframed what an individual developer could produce. What Endava is describing is the organisation-level version of that shift, applied to a whole delivery business at once.
Indian Angle
Nowhere does this land harder than in India. The country's technology services industry exports well over $200bn a year and was built on the same headcount-linked, time-and-materials model that agents now threaten to compress. The giants - Tata Consultancy Services, Infosys, Wipro, HCLTech, Tech Mahindra and LTIMindtree - are chasing the same answer Endava is, and a near-shore rival moving first only raises the pressure.
The Indian majors are not standing still. Platforms such as Infosys's Topaz and Wipro's ai360 are pitched at exactly this agentic future, and most have struck partnerships with the large model labs. The harder consequence is on hiring. The big firms have already slowed campus recruitment, and a delivery model that needs fewer junior engineers to produce the same output puts the classic pyramid, with lots of freshers at the base, under real strain.
There is a governance dimension too. Running tools like ChatGPT Enterprise over client code raises data-residency and confidentiality questions that Indian providers must answer under the Digital Personal Data Protection Act and increasingly demanding client contracts. For India's millions of engineers, the upside is leverage: the same talent base, equipped with agents, could move up the value chain from writing code to designing and supervising the systems that write it.
FAQ
What exactly is Endava doing?
Endava says it is adopting ChatGPT Enterprise across its workforce and using OpenAI's Codex to automate parts of software delivery, redesigning workflows around AI agents to build what it calls an "AI-native" culture rather than treating the tools as a simple add-on.
Why should Indian IT firms care?
Endava competes for the same enterprise software work using a similar delivery model. A rival reorganising around agents pressures Indian majors to prove their own platforms can protect margins as the link between headcount and revenue weakens.
Does this mean fewer tech jobs in India?
Not necessarily fewer, but different. Routine coding at the base of the pyramid is most exposed, while engineers who can direct, review and integrate agent output become more valuable.
Where can I read the original?
OpenAI published the full account on its website, linked below.
This story was reported by OpenAI. Read the full original coverage at OpenAI.