New Delhi, April 27 -- As AI adoption matures across India Inc, the conversation is shifting from pilots to production-and from hype to measurable outcomes. In an interview with TechCircle, Rishi Aurora, managing partner at IBM Consulting for India and South Asia, outlines what separates real transformation from experimentation, the execution gaps holding enterprises back, and why digital sovereignty is emerging as a core design principle in enterprise AI. Edited excerpts:
India Inc appears to have moved beyond the AI "hype cycle". What distinguishes meaningful AI-led transformation from costly experimentation?
The gap today is between activity and autonomy. Many organisations are still running AI pilots that sit outside core workflows and enterprise control-these are expensive experiments rather than a transformation. Our research shows nearly 79% of executives expect AI to contribute significantly to revenue this decade, but only 24% have clarity on where that value will come from. That gap reflects ambition without execution discipline.
Meaningful transformation happens when AI is embedded into how work actually runs-decision-making, execution, and outcomes-while maintaining visibility over data, models, and risk. As AI moves closer to sensitive data, digital sovereignty becomes central. Enterprises are redesigning workflows so AI operates where data resides, across hybrid environments with governance and accountability built in. At IBM, through our "Client Zero" journey, we have reinvented more than 70 workflows using AI and automation, driving $4.5 billion in productivity. When AI becomes part of business architecture-integrated by default and sovereign by design-it evolves into a durable growth engine.
As enterprises shift from pilots to production, what are the biggest bottlenecks in scaling GenAI and agentic AI?
The biggest constraint isn't AI-it's enterprise readiness. We consistently see four bottlenecks: legacy processes not designed for AI, high technical debt, fragmented data, and a shortage of skills. Governance is another underestimated challenge. As AI becomes more autonomous, explainability and oversight must be designed in from day one.
Scaling GenAI and agentic AI requires simultaneous change-modernising systems, strengthening data foundations, redesigning workflows, and preparing the workforce. Addressing these in isolation doesn't work. That's why many pilots fail to translate into enterprise-scale value.
Agentic AI promises autonomous decision-making. What risks are enterprises managing-and where are they underestimating exposure? With regulations like the Digital Personal Data Protection Act, enterprises are rightly focused on privacy, compliance, and explainability. However, risks are often underestimated in integration and governance. AI agents don't fail in isolation-they fail in fragmented systems without clear accountability. In sectors like banking, where AI is used in loan origination or KYC workflows, a single flawed decision can quickly multiply risk.
Security is another blind spot. According to IBM's X-Force Threat Intelligence Index 2026, attackers are using AI to exploit vulnerabilities faster than ever. As AI accelerates decision-making, it also accelerates exposure. The real issue isn't autonomy-it's autonomy without transparency, lineage, and human oversight.
Which sectors in India are leading AI adoption at scale, and what lessons are emerging?
Banking, telecom, manufacturing, and government are clearly ahead. Banks have scaled AI with strong governance and human-in-the-loop systems, while manufacturing is advancing in real-time optimisation and agentic operations. Government initiatives like BharatGen show how AI can scale with trust and inclusion.
The key lesson isn't sector-specific-it's transferable. Banking's governance discipline can inform healthcare, while manufacturing's models can reshape supply chains. Across industries, one principle holds: AI delivers compounding value only when embedded end-to-end into workflows, supported by high-quality data and governance.
Digital sovereignty is becoming central. How is it reshaping enterprise architecture?
Digital sovereignty is no longer a compliance requirement-it's a strategic advantage. It underpins trust, resilience, and long-term competitiveness. Enterprises are now designing AI systems with sovereignty at the core. In sectors like banking, telecom, and healthcare, control over data and models is non-negotiable. This is accelerating hybrid cloud adoption-blending public, private, and sovereign environments.
The priorities are clear: design for sovereignty upfront, ensure AI operates where data resides, and transform entire workflows-not just applications.
How should CIOs balance legacy modernisation with AI adoption?
Layering AI on broken foundations rarely works. The goal isn't to rebuild everything, but to prioritise what matters.
Leading organisations focus on improving how work flows across systems-so data, decisions, and actions connect seamlessly. This creates a stable base for AI adoption without disrupting operations.
Many are extending the life of existing platforms by making them more flexible, rather than replacing them entirely. The objective is readiness-systems that support both current operations and future AI-driven scale.
As AI gets embedded in decision-making, how are organisations redefining accountability and workforce roles?
There's a clear shift underway. Leaders are moving from asking "Did the system work?" to "Who owns the outcome?" AI is increasingly handling analysis and execution, but accountability remains human. Leadership teams are defining guardrails, escalation paths, and points where human judgment must intervene.
This is creating an augmented workforce-where employees work alongside AI agents. Roles are shifting from execution to oversight and decision-making. Organisations investing in reskilling are seeing faster adoption and greater trust.
What role do consulting partners play as enterprises become more AI-mature?
The expectation has changed. Enterprises no longer want just advice-they want partners who co-innovate, co-own outcomes, and stay through implementation. At IBM, we combine consulting expertise with enterprise-grade platforms to help clients scale AI safely. Solutions like IBM Enterprise Advantage and Sovereign Core enable organisations to build, govern, and operate AI with control and consistency.
Equally important is credibility. Through our own "Client Zero" journey, we've applied AI across our workflows first-understanding where value compounds and where risks emerge. That allows us to offer clients a practical blueprint, not just theory.
Published by HT Digital Content Services with permission from TechCircle.