Reclaiming Control in the AI Era: A Conversation with Kalyan Kumar, CPO, HCLSoftware

For more than a decade, cloud adoption shaped digital transformation across industries, with public cloud and SaaS emerging as the fastest path to scale and a pathway to modernization. Today, as artificial intelligence becomes embedded in business operations, public infrastructure, and high-stakes decision systems, the conversation is shifting. Enterprises are now questioning what they should control, what they should consume, and what must remain sovereign.

This rethink is unfolding alongside a broader debate about data residency, AI governance, and pivoting a reimagining of digital infrastructure. As regulations evolve and geopolitical considerations influence technology choices, organisations are reassessing the long-standing assumption that everything must live on public cloud platforms.

At the centre of this shift is HCLSoftware, the $1.4 billion enterprise software division of HCLTech, serving more than 20,000 customers globally. The company positions its intellectual property model as built in India for the world, with engineering distributed globally but IP ownership and registration anchored in India. HCLSoftware is the largest enterprise software company headquartered out of India globally.

Driving this strategy is Kalyan Kumar, Chief Product Officer of HCLSoftware and his team. With more than two decades at HCL, he has held leadership roles across solutions, alliances, business development, digital practices, product engineering and delivery. His career spans early work in Infrastructure Management Services starting in 2000, leading the Digital Foundation Practice during a period of peak growth, forming the Intelligent Operations AI software division, and contributing to one of the industry’s early business service management SaaS platforms in 2008. Today, as CPO, he leads product strategy, product specialists, product management, engineering, professional services and operations for HCLSoftware.

In this discussion with Dataquest, he shares his perspective on digital sovereignty, the emerging concept of AI repatriation, how enterprises are approaching private AI stacks, and why flexibility and choice are beginning to replace the cloud-first doctrine.

You used the term ‘Great AI Repatriation.’ How do you define it in the context of enterprise tech evolution?

The shift happening now is broader than events like cloud repatriation. Earlier, organisations moved to the cloud because it was the fastest way to transform. What I call the Great AI Repatriation is about taking back control across the AI stack. This is not only about where data sits, but who owns the IP, who controls access and who governs continuity.

Engineering talent will always be distributed. We have teams in the US, Israel, Rome, Australia, Manila and a bulk of  our engineering teams in India. But the intellectual property we build is owned in India. It does not matter where in the world we sell the product. The ultimate beneficiary of that IP is India. That mindset is key to how we look at ownership and sovereignty.

Many organisations mix up terms such as sovereignty, localisation and residency. How do you explain the difference?

There is a lot of confusion. Data residency is about where data physically lives. Some countries mandate that certain kinds of data cannot leave their borders. That is residency.

Sovereignty is much broader. It is about choice and control. Even if your data sits in your country, if someone else can shut off access to the platform or service you depend on, you are not sovereign.

There are four dimensions we look at:

• IP sovereignty
• Data sovereignty
• Technology sovereignty
• Commercial sovereignty

Commercial sovereignty is often overlooked. If you fully depend on subscription access and someone decides to disconnect it, your business can stop. We saw a large global oil company face this recently. They owned their data, but they did not control the platform. That is the gap.

How is this influencing the way enterprises are designing AI and data platforms?

Enterprises are not rejecting cloud, but they are designing more hybrid strategies. Public models are useful for commodity or open use cases. But when it comes to proprietary insights, private data and critical workloads, organisations want private AI stacks.

To do that, companies are building multimodal databases, vector databases, metadata governance layers and modern enterprise data stacks. They are experimenting with open models, small language models and model compression. Retrieval augmentation is also becoming important to tune and apply these capabilities.

So there is a shift happening. Earlier everything was about cloud-first and SaaS-first. Now customers want flexibility and long-term control.

HCLSoftware has introduced an XDO blueprint. How does it support this shift?

XDO stands for Experience, Data and Operations. It is a method, not a product. We use it as a blueprint to engage customers and partners.

Every organisation begins with experience. That could be customer experience, employee experience, partner experience or product experience. The next layer is data: how it is organised, governed and used to create intelligence. The third layer is operations, where the goal is efficiency, agility and lifecycle management.

We also map our products into this blueprint. Customers can use it to align their roadmap with sovereignty, residency and control requirements. It helps avoid decisions that are convenient today but restrictive later.

Which markets and industries are showing the strongest momentum for this approach?

We operate in 132 countries directly and through partners. Adoption is strongest in government, public sector, financial services, healthcare and manufacturing. Retail and consumer businesses are also becoming more aware that technology choices today impact resilience tomorrow.

If you had to give one piece of guidance to a CXO building an AI operating model for the next phase of digital transformation, what would it be?

Start with clarity. Map your experience, data and operational needs. Decide what you want to own and what can be consumed. Understand data residency requirements and cross-border data flows based on your operating footprint.

Technology exists to drive business value. In some cases, technology is the business. Convenience in the short term can make scaling, integrating and complying much more complex later. A blueprint approach helps prevent that.

 Looking toward 2026, which strategic priorities will define the next phase for HCLSoftware?

Our priority is to continue developing the XDO blueprint. We are bringing together the key elements needed for deployment choice, enterprise orchestration, AI-driven experience and data intelligence across distributed environments. The focus is enabling customers to run workloads wherever they choose and retain full control.

As AI becomes a core part of enterprise intelligence and national infrastructure, the questions are no longer just technical. They are organisational, regulatory and strategic. The next wave of enterprise AI adoption will likely be shaped by control, flexibility and sovereignty- not only speed.