The Indian IT services industry has stepped into FY27 with a concerted mission: stripping away legacy technology debt. Major tech firms, including Tech Mahindra, have established aggressive internal benchmarks, targeting a remarkable 30 to 35 percent reduction in technical debt for their enterprise clients. Yet, despite successfully modernizing nearly 90 legacy applications and achieving 60 percent faster migration cycles through AI-enabled software development, a profound paradox plagues corporate boardrooms. Enterprises are hesitating, gripped by a singular anxiety: that the cure—artificial intelligence—might ultimately incubate a far more complex disease.
Conversations with industry leaders across the manufacturing and BFSI sectors reveal deep-seated apprehension. While IT providers are aggressively pivoting toward AI-led deal wins, clients worry about inadvertently engineering a new variant of technical debt. Vijay Balakrishnan, Chief Digital & Information Officer at Godrej Enterprises Group, encapsulates this unease regarding AI-assisted coding. Without complete organizational visibility into every line of generated code, companies risk building invisible vulnerabilities. Balakrishnan aptly notes that technical debt operates much like the law of conservation—it cannot be destroyed, only transformed from one form into another.
This sentiment resonates strongly within the financial services ecosystem. Banking, Financial Services, and Insurance (BFSI) institutions are pumping the brakes due to three critical roadblocks: the inherently probabilistic nature of AI that hinders large-scale production deployment, persistent regulatory grey areas, and the absence of robust governance frameworks. Furthermore, the spiraling token-based cost of AI usage has transformed into a major balance sheet liability. Sanjay Varma, President of the Fintech Solutions Group at Aurionpro Solutions, warns that unmanaged compute costs can quickly turn AI-driven digitization into a severe financial drain, urging firms to prioritize token efficiency.
This cautious stance is visibly rippling through the broader macroeconomic landscape, heavily influencing the performance outlook of tech heavyweights like Tata Consultancy Services (TCS) and Wipro. During recent earnings calls, Wipro CEO and Managing Director Srini Pallia acknowledged that discretionary technology spending is facing downward pressure as enterprises reallocate budgets toward AI integration, squeezing traditional IT services and support functions.
Experts point out that the structural nature of AI debt makes it fundamentally different from historical legacy issues. Titus M, Practice Director at Everest Group, points out that AI injects exponential variables into the technology stack. Unlike linear legacy debt, a single suboptimal decision regarding AI model selection, weak data foundations, or inefficient token usage can trigger disproportionately higher financial and operational costs. However, Titus remains optimistic, noting that AI-related technical debt can be actively remediated in real time through outcome-based operating models and AI FinOps frameworks.
Ultimately, Indian enterprises are walking a tightrope. While they continue to monitor adoption risks, AI is steadily weaving itself into enterprise architecture without demanding sweeping, capital-intensive infrastructure overhauls. As organizations mature in their AI journeys, the initial spike in perceived debt is expected to smooth out, paving the way for long-term operational efficiency.
"The paradox of artificial intelligence adoption in enterprise ecosystems highlights a critical maturation phase for the Indian technology landscape. While cutting legacy tech debt remains an urgent operational priority, business leaders are rightly concerned about trading predictable old liabilities for compounding, exponential new risks driven by probabilistic algorithms and unmanaged token economics. At StartupLanes, we observe that forward-thinking enterprises are no longer rushing blindly into AI deployments. Instead, they are demanding rigorous outcome-based operating models, robust data foundations, and strict AI FinOps governance. As we navigate FY27, the true winners in the digital transformation race will not be those who adopt AI the fastest, but those who successfully balance speed with sustainable cost predictability and long-term architectural integrity." — Dr. Shishir Gupta, Founder & CEO, StartupLanes
