The biggest barrier to scaling generative AI is not the model itself; but the enterprise environment around it. Technically, many organizations struggle with fragmented data, weak metadata, inconsistent governance, and systems that were never designed to support AI at scale. Niraj Kumar, Chief Technology Officer, Onix exclusively informs Bhavya Bagga, Business Reporter, APAC Media & CXO Media that Onix through its Wingspan platform helps enterprises modernize faster, reduce complexity, and operationalize AI in ways that are context-aware, secure, and aligned to real business objectives.
You have led technology innovation across cloud, data, and AI for nearly two decades. As CTO of Onix, what is your vision for helping enterprises move from traditional cloud transformation to AI-first business transformation, and what role does Onix play in this journey?
As CTO of Onix, my vision is to help enterprises move beyond cloud as an infrastructure milestone and toward AI as a business operating model. Cloud transformation gave organizations scalability and agility, but AI-first transformation is about using intelligence to continuously optimize decisions, workflows, and outcomes across the enterprise. The real shift happens when data, cloud, and AI are no longer treated as separate initiatives but as a unified foundation for business reinvention.
At Onix, we play the role of an enabler and accelerator in that journey. Through our Wingspan platform and our broader cloud, data, and AI capabilities, we help enterprises modernize faster, reduce complexity, and operationalize AI in ways that are context-aware, secure, and aligned to real business objectives. Our focus is not just on building AI solutions but on creating the intelligent foundation enterprises need to scale them with confidence.
Many enterprises have experimented with generative AI but struggle to move beyond pilot projects. In your view, what are the biggest technical and organizational barriers to scaling AI into core business operations, and how can organizations overcome them?
The biggest barrier to scaling generative AI is not the model itself; it is the enterprise environment around it. Technically, many organizations struggle with fragmented data, weak metadata, inconsistent governance, and systems that were never designed to support AI at scale. Organizationally, the challenge is often that AI remains isolated in innovation teams or pilot programs, without clear ownership from business and operations leaders.
To overcome this, enterprises need to shift from experimentation to architecture. That means defining strong data foundations, creating reusable AI patterns, embedding governance from the start, and aligning AI initiatives to measurable business outcomes. Equally important is change management: teams need to understand how AI augments their work, what guardrails exist, and how success will be measured. When AI is tied to business processes rather than isolated use cases, it becomes much easier to scale.
Agentic AI is rapidly evolving from simple assistants to autonomous systems capable of making decisions and executing tasks. How do you see this changing enterprise workflows, and what governance frameworks are essential to ensure trust, accountability, and security?
Agentic AI will fundamentally change enterprise workflows by moving AI from a reactive assistant role to an active execution layer. Instead of simply answering questions or generating content, agents will increasingly orchestrate multi-step tasks, coordinate across systems, and support decision-making in real time. That means workflows will become more dynamic, more automated, and far less dependent on manual intervention for routine execution.
With that shift, governance becomes non-negotiable. Enterprises need frameworks that define clear human oversight, role-based permissions, auditability, model transparency, and policy controls for autonomous actions. Security must extend beyond model access to include data lineage, task execution boundaries, and exception handling. The goal is to create a trusted operating model where autonomy is introduced intentionally, with accountability built into every layer.
AI is only as effective as the data and business context behind it. How can enterprises build intelligent data ecosystems that enable AI to deliver accurate, context-aware insights rather than just generating outputs?
AI becomes truly valuable when it understands business context, not just raw data. To build intelligent data ecosystems, enterprises need more than data lakes or warehouses; they need a semantic layer that connects data to business meaning, relationships, and intent. That includes strong master data practices, metadata management, lineage visibility, and a shared business vocabulary that AI systems can use to interpret information correctly.
This is where context-aware architecture matters. When data is enriched with business rules, domain knowledge, and system dependencies, AI can generate insights that are more accurate, relevant, and actionable. At Onix, we believe the future lies in living data ecosystems that continuously reflect enterprise reality, rather than static repositories that only store information. That is what enables AI to move from producing outputs to driving decisions.
Looking ahead, what emerging AI technologies or enterprise trends do you believe will have the greatest impact over the next three to five years, and how should CIOs and technology leaders prepare for this next phase of AI-driven transformation?
Over the next three to five years, I believe the most important shifts will come from agentic AI, multi-agent orchestration, semantic intelligence, and the deeper integration of AI into core enterprise systems. We will also see synthetic data, domain-specific models, and AI governance platforms become much more central to enterprise adoption. The real differentiator will not be who has access to AI, but who can operationalize it responsibly and at scale.
CIOs and technology leaders should prepare by investing in data modernization, establishing governance early, and building cross-functional teams that combine business knowledge with technical expertise. They should also focus on creating flexible architectures that can adapt as AI capabilities evolve. The organizations that win will be the ones that treat AI not as a standalone tool, but as a strategic capability woven into the fabric of the business.
