The idea that AI will somehow make Indian IT firms irrelevant keeps resurfacing every few months. But Infosys’ recent partnership with Cognition offers a more grounded reality.
Infosys is collaborating with Cognition to deploy Devin, described as the world’s first AI software engineer, across its internal engineering teams and client engagements, also through its composable AI agents platform, Topaz Fabric. The IT giant had already been using Devin internally, recognising improvements in engineering quality and efficiency. The rollout now moves from experimentation to scaled delivery.
The goal is to speed up software development, reduce time-to-market, and automate slow and manual engineering.
That’s where the story gets interesting. It signals that AI companies need Indian IT firms to deploy their product.
What Does it Mean for Indian IT?
Infosys plans to embed Devin within internal workflows, deploy it into client delivery models, and enable customers to run it inside their own environments. Scott Wu, founder and CEO of Cognition, framed the partnership as a way to bring autonomous engineering into the heart of complex enterprises.
“Silicon Valley tech is often too insular and self-referential; building for ourselves is easy, but solving the messy problems to meet customers where they are is hard,” Wu posted on X.
The timing of the partnership is also notable since it follows Infosys’ AI alliance with AWS, which aims to bring together Infosys Topaz and Amazon Q Developer to improve software delivery and internal operations.
A clear pattern is emerging. Native AI companies are building powerful tools, cloud providers are embedding them into their platforms, and Indian IT firms are becoming the execution layer.
In a conversation with AIM, Gaurav Vasu, CEO of the cognitive intelligence platform UnearthInsight, observed a familiar cycle repeating. “The partnership between Infosys and Cognition AI signals that AI-era alliances will closely mirror earlier product-service ecosystems built around SAP, Microsoft, Oracle, and IBM,” he said.
Earlier, enterprise software vendors depended on system integrators to modernise global businesses at scale. AI platforms, Vasu argued, will follow the same path.
This also challenges the idea that AI will hollow out Indian IT. Vasu noted that AI is changing the nature of work rather than eliminating the role of service firms. Coding agents can generate and refactor code, yet large transformation programmes still require integration, security, compliance, and change management.
An AI agent can rewrite an application. Deploying it across a global bank still needs domain experts, architects, and delivery teams.
That shift has implications for hiring. Routine tasks will shrink. Demand will grow for AI engineers, data specialists, platform architects, and governance roles. “Overall, AI improves productivity and speed of delivery, while Indian IT firms remain the enterprise execution layer that turns AI capability into real business impact,” Vasu added.
Sanchit Vir Gogia, founder and CEO of Greyhound Research, predicted an impact on hiring. “This is not the end of Indian IT hiring, but it is absolutely the beginning of a shift in what those hiring patterns look like,” he told AIM.
He explained that tools like Devin change the economics of work. “It reduces the need for large pools of entry-level engineers who were previously tasked with repetitive, rules-based tasks. Those roles are now exposed to automation.”
At the same time, he noted that demand is moving up the stack. “Firms are prioritising engineers who can supervise AI agents, design workflows that include them, and ensure compliance and reliability.”
He also pointed to client pressure for optimising deliverables. “Enterprise buyers are now expecting measurable productivity gains, and they expect those to show up in either reduced pricing or accelerated timelines.” This is already affecting staffing models. “Providers are being asked to demonstrate how they will deliver more with less, and that has a direct impact on workforce planning,” he added.
The result, Gogia remarked, is a reshaping of the classic IT pyramid. “The wide base of junior staff is narrowing, and the focus is shifting to mid-tier and senior roles with deeper skills in orchestration, governance, and AI oversight.”
The Infosys-Cognition deal also fits into a wider pattern across the industry. Cognizant has emerged as a major revenue partner for Lovable. Accenture has partnered with Bolt. Hexaware recently tied up with Replit. These alliances point to agentic AI becoming a new business vertical for Indian IT.
What Does it Mean for Infosys?
Infosys is hedging against disruption; it is rather positioning itself as a conduit for enterprise AI.
Historically, Vasu noted that ecosystem partnerships contributed 40–50% of revenue for Indian IT firms. Agentic AI partnerships are still small in comparison, estimated at 5–8%of revenue, yet their strategic importance is high.
Gogia affirmed that the Infosys-Cognition deal marks a clear shift from experimentation to execution. “This is not a vendor alliance for narrative gain or tooling access. What we are seeing here is a calculated move to embed autonomous agents into the core engineering ecosystem, both internally and within client-facing projects.”
“It is no longer about how many engineers you can staff, or how many locations you can deliver from. It is about how fast, safely, and predictably you can deploy change across large and messy technology estates,” he added.
Infosys is signalling that AI agents will become core delivery building blocks.
“The industry has historically been anchored to effort-based pricing models. In an agent-enabled future, that model loses power,” Gogia remarked, adding that Infosys is trying to move toward outcome-led conversations. “The use of agentic AI tools such as Devin is being framed not as a productivity hack, but as a rethinking of how software is engineered and delivered.”
The deal is giving the wider industry a reality check. “These partnerships will act as a litmus test for existing internal agent programmes,” he noted, signalling that IT firms will now have to evaluate their homegrown agents against platform-grade agents like Devin on auditability, performance, and integration.
In the last quarter, Infosys reported strong results, with profits jumping 13.2% year-on-year to ₹7,364 crore, but stopped short of giving a breakdown of AI revenue. That contrasts with peers like TCS and HCLTech, which have already put a number on AI contributions.
In the end, AI tools, no matter how advanced, still need organisations that understand scale, regulation, legacy systems, and business risk. With more IT firms poised to collaborate with native AI companies, the apocalyptic future of AI engineers replacing humans may never appear.
The post Infosys–Cognition Deal Signals Why AI Cannot Skip Indian IT appeared first on Analytics India Magazine.
