BENGALURU: The IT industry is witnessing one of the deepest structural transformations in decades as artificial intelligence (AI) reshapes talent needs across the delivery chain. The traditional hierarchical model of large teams of junior programmers overseen by layers of managers is giving way to a capacity-dense, AI-enhanced architecture.
“There has been a shift from depth of coding to problem framing,” said Gilroy Matthew, UST’s chief operating officer. Instead of testing whether candidates can write a Java function, companies are now asking whether they can identify business problems, break them down into actionable steps with AI, and validate the output for bias, risk, and completeness. Access to AI tools narrows the gap between new engineers and mid-level engineers. A well-trained entry-level employee is expected to become productive faster than before.
.
LTIMindtree describes the emerging structure as a “diamond” rather than a pyramid. At the base is a smaller layer of engineers who are proficient in AI, supported by automation and AI agents that handle routine execution. In between, architects and managers evolve into orchestrators who design AI-first workflows, integrate systems, and align delivery with business goals. “It’s not about fewer people; it’s about fewer people doing low-discretionary work,” said Gururaj Deshpande, chief delivery officer at LTIMindtree, arguing that AI-led productivity gains will help clear larger backlogs of projects rather than simply reducing headcount.
One obvious shift is the shrinking space for purely supervisory managers. In an AI-led environment, managers will be expected to redesign processes around automation, understand agent-based workflows, ensure AI governance and link implementation to business KPIs.
Employment models are changing in parallel. At UST, screening has moved from “employability” to “adaptability.” Candidates are given realistic business scenarios and are evaluated on how they frame problems, use AI tools, validate outputs, and consider ethical considerations.