For decades, the economics of IT services were relatively straightforward: more projects meant more people, and more revenue generally meant a proportional increase in headcount. AI is beginning to disrupt that equation.
At Coforge, the change is already visible in the numbers. Over the last year, the company grew revenue by about 30% while headcount increased 10%—a divergence that Sunil Fernandes, COO & EVP, Coforge, describes as a significant break from the traditional linearity of the industry.
But Fernandes argues that the bigger story is not productivity alone. Coforge is attempting to redesign the operating model itself—how it hires, trains, sells, delivers, builds software, and structures teams.
“The quantum of change and the pace of change have gone up tremendously,” he says, describing the past year as a period of “complete reinvention.”
That reinvention is increasingly showing up in Coforge’s business. In Q1 FY27, 86% of revenue came from AI-led engineering, data and cloud services, while approximately 30% of active engagements were leveraging AI within delivery workflows. Across those workflows, the company reports 25%–35% productivity gains in AI-assisted code generation and review, and 40%–60% gains in intelligent legacy modernisation. Modernisation timelines have been compressed by roughly 10x, with migration becoming approximately 5x faster.
The implication is significant: AI is not simply becoming another service capability. It is changing the relationship between revenue, effort, expertise and time.
Beyond the pyramid
The first place Coforge is redesigning that relationship is its workforce.
Traditional IT services were built around a pyramid: large numbers of engineers at the base, progressing through multiple hierarchical levels. Coforge’s previous architecture had 13 levels. Its new model reduces that to four levels across eight career tracks, with greater emphasis on expertise rather than layers of hierarchy.
That is not merely an HR restructuring. It reflects a bigger change in the unit of work.
Coforge is taking its approximately 46,000 employees through the transition, combining internal reskilling with external hiring for AI engineers and architects.
Fernandes’ description of the new workforce is revealing: the employee is moving from someone who does the work to someone who steers the agent doing the work.
“The developer who’s writing Java code” is no longer necessarily the person producing the code, he explains. The agent does that work; the engineer guides it and ensures the output meets the requirement.
That distinction goes to the heart of Coforge’s AI strategy.
The rise of the Mod Squad
The company’s Mod Squads bring humans and AI agents together in delivery teams. The model is being rolled out across 40+ client accounts, with 50+ programs operating under 30-, 60-, and 90-day transition plans. Across client technology environments, more than 150 AI agents are live or in build.
The agents can be specialised around business domains such as claims and underwriting, or technology functions such as modernisation, SRE, data engineering and infrastructure operations.
The significance goes beyond automation. It changes what Coforge believes a delivery team should look like. Its ambition is for new engagements to increasingly run through an agentic delivery platform, rather than treating AI as an add-on to an otherwise conventional services model.
That is also why the company’s productivity metrics matter. A 25%–35% uplift in AI-assisted coding and review, alongside 40%–60% improvement in intelligent legacy modernisation, is valuable not simply because fewer hours may be required, but because clients can bring capabilities to market sooner.
Fernandes puts the emphasis on time-to-market, saying Coforge has typically achieved at least 20% acceleration in delivering business capabilities.
AI is the means, not the destination
This distinction is central to Coforge’s technology strategy.
The company’s Nuuron enablement suite is not positioned simply as an AI platform. It combines processes, approaches, templates, and technology assets designed to create what Fernandes calls an intelligence layer across the enterprise.
“The headline is we basically say enabling enterprise autonomy is our goal,” he says. “Our goal is not AI. AI is a means.”
Coforge says more than 160 clients are working with Nuuron to enable their AgenticOps journey. Behind it sits a growing portfolio: 11,000+ data and AI practitioners, eight AI platforms, 22 AI assets, and more than 100 reusable AI agents and accelerators.
The strategy is to encode Coforge’s accumulated engineering and domain expertise into reusable technology. Its platforms incorporate knowledge from areas including travel, insurance, and banking; the company, for example, has experience across 65+ airlines globally.
The resulting proposition is less about selling an AI tool and more about compressing the journey from problem identification to enterprise-scale execution.
Reimagining before automating
That approach also explains why Fernandes is sceptical of enterprises rushing from AI experimentation to isolated use cases.
He argues that companies must first rethink their processes, data foundations, talent and governance. AI applied to an inefficient legacy process may simply automate the inefficiency.
“You have to be able to reimagine processes first before you start ladling AI on top,” he says. Governance, trust and security, he adds, cannot be bolted on later.
Coforge has attempted to apply the same discipline internally. The company first had to organise its own data, establish context layers and build governance and security before scaling AI across internal functions.
That internal experience is becoming part of its client proposition.
From efficiency engine to reinvention engine
The company’s transformation is also being shaped by large-scale client engagements. Fernandes cites a $1.5 billion transformation with a global travel technology services firm, where an AI-first approach spans talent induction, training, legacy-system understanding and modernisation.
The acquisition of Encora further strengthened Coforge’s AI engineering capabilities and expanded its industry footprint.
For Fernandes, however, the ultimate transformation is organisational. The CEO’s mandate, he argues, is shifting from running the business-as-usual execution engine to “completely reimagining the organisation in every dimension” and becoming the primary driver of change.
That is perhaps the most consequential shift in Coforge’s AI journey. The question is no longer how many people an IT services company needs to deliver a given amount of work. It is how much intelligence, expertise, and business capability it can orchestrate through a combination of people, platforms, and agents.
And Fernandes frames the challenge bluntly: “If you can convincingly tell clients that you can drive enterprise autonomy for them, just drive enterprise autonomy for yourself.”
For Coforge, that is no longer a future-state ambition. It is becoming the operating model.