NTT DATA’s Subhashini Desigan explains how enterprises are moving from AI pilots to scalable, agentic AI models by combining modernisation, digital twins, workforce transformation, and outcome-driven business strategies.
As enterprises move deeper into the AI era, businesses are now under pressure to operationalise AI at scale while balancing governance, infrastructure modernisation, technical debt, workforce transformation, and measurable returns. At the same time, service providers are restructuring their portfolios around agentic AI, intelligent automation, and AI-native operating models to meet changing enterprise expectations.
In this interaction with CIO&Leader, Subhashini Desigan, Executive Managing Director – Apps & BPS (North America) and Global Strategy & Operations Head, NTT DATA, discusses how the organisation is embedding AI across applications, business process services, data modernisation, and enterprise operations. She also shares insights into NTT DATA’s agentic AI strategy, digital twins, AI Centers of Excellence, operational transformation models, AI-first leadership, and the growing importance of ecosystem partnerships in building scalable AI-native enterprises.
CIO&Leader: You have managed over a billion-dollar North America P&L. To get an overview, how are you operationalising the shift from traditional effort-based delivery to an outcome-led, AI-driven business model?
Subhashini Desigan: Just talking about our overall business, I have a global remit as the Global Strategy and Operations Leader. In North America specifically, I run the business. We have been on this journey of moving toward AI for a few years now, and GenAI came up a few years ago, which already feels like a lifetime.
I will talk about it in three phases — our offerings and solutions, and then the human aspect of how we are skilling our workforce.
From an offering standpoint, we have our traditional offerings in applications around the software engineering lifecycle, quality engineering, and application management. All of these offerings are now infused with AI, specifically agentic AI. We have accelerators, tools, and platforms around each of these, including our own aXet platform to accelerate the software development lifecycle. For example, we are working with a large high-tech company to completely reimagine their SDLC processes.
The same applies to Business Process Services. We do BPS processing in insurance and health plan spaces. Earlier, we used bots, and we still continue to have some, but we also have patented technology to convert bots into agents. Now we have converted those BPO processes into agentic models as well.
On the data side, our data migration, modernisation, and visualisation capabilities are all infused with agentic AI. We also have agentic data products and cloud migration capabilities. Across migration and modernisation, we use agentic AI-based accelerators and partnerships with startups. So, all of our traditional offerings are now infused with AI and agentic AI.
We also have newer offerings specifically for this AI era, such as agentic BPaaS across industries including insurance, healthcare, banking, manufacturing, logistics, and supply chain operations. In addition, if you look at technical debt modernisation, we have always done cloud modernisation, but now it is infused with agentic AI. We also have newer offerings focused on how to work with legacy systems while reducing technical debt and accelerating modernisation.
For this, we use digital twin technologies. We create digital twins of the business process layer, including non-functional elements, which help with real-world simulation, predictive analytics, dependency management, and faster enhancements. This allows us to reduce technical debt by adding an agentic layer on top.
We also have digital twins on the data side. This is foundational and is changing the game because it helps democratise data using AI and semantic ontologies.
Lastly, we have instituted AI Vista. We, as the Applications, BPS, and Data practice, work closely with AI Vista as well as our AI Go-To-Market unit. Our CEO, Abhijit Dubey, has taken on the central role as Chief AI Officer. It goes to show that AI is never siloed or secondary, but embedded into every strategic decision driving growth and transformation.
AI is never siloed, never secondary, but always embedded in every strategic decision that drives our growth and transformation.
On the people side, this is also a massive transformation. We have established Centers of Excellence and training programs to reskill our workforce. It is also a cultural transformation because people need to continuously challenge themselves to use AI with customers.
We now have Centers of Excellence focused on OpenAI, Mistral AI, hyperscaler technologies, and our own aXet platform. AI is a major focus area for us in terms of people, training, and building a future-ready culture.
Lastly, there are ongoing conversations with customers as part of this transformation journey.
CIO&Leader: NTT DATA has a goal to reach US $2 billion in revenue through AI agent solutions by 2027. What are the specific portfolio shifts within the application and BPS business that you are prioritising to capture market share right now?
Subhashini Desigan: We are looking at transformed offerings, which are AI-enabled. We are looking at new offerings that I talked about around agentic BPaaS, technical debt modernisation, leveraging semantic ontology layers, and AI-native businesses. Those are the three levels we are looking at to drive agentic AI and AI business.
Again, NTT DATA, especially from the Applications, BPS, and Data practice, is very domain-centric. We do not look at AI as a standalone technology. AI helps us drive better business outcomes. That is how we approach it in our practice, focusing on what business outcomes we can accelerate for clients and what new outcomes we can help them achieve that were not possible earlier.
We do not look at AI as a standalone technology. AI helps us drive better business outcomes.
These are the conversations we are having.
We are confident about the US $2 billion target because, first, we are well positioned with full-stack capabilities to deliver holistic solutions that accelerate business outcomes. Also, because we have data centers, connectivity, infrastructure, and more, we are able to build strong partnerships with customers to drive this change.
For example, we can take on some of their technical debt through our data center capabilities and free up their operating expenses, allowing them to invest faster in AI. We also bring innovative commercial models to support this journey.
This helps us stay closely aligned with customers as partners, and we are already doing this with many of our accounts, including large manufacturing clients we have signed over the past year.
CIO&Leader: Enterprises are shifting focus from AI pilots to measurable business outcomes. What would be your recommendation for CXOs trying to balance long-term AI-native infrastructure investments with immediate returns?
Subhashini Desigan: We are looking at it in a few ways. The timing is actually perfect because we just concluded our client advisory council two weeks ago, where some of our key client advisors spent two days discussing what they are thinking, what pressures they are facing, and how we can help them navigate those challenges.
It is very clear that CIOs and business leaders now have to show ROI on the investments they started a few years ago. This also means that in many areas, they need to move from POCs or proofs of value to enterprise-grade implementations.
The foundational layer of the agentic data value chain, creating the infrastructure and partnerships to support it, is non-negotiable. Most enterprise customers need to continue investing in these foundational elements, and many have already started that journey.
Another area we are focusing on is operations. While IT spend is significant, operational spend is much higher. We are working with customers to transform their operations using AI so they can see immediate benefits. This helps free up funds that can then be reinvested into priority areas.
Focus is critical. We work with each customer to identify their highest-priority areas. For some, it is launching new business lines or getting to market faster with AI. For others, such as insurance companies, it is about becoming agentic insurers because early movers will have a strong competitive advantage.
Priorities vary, but focusing on a few areas that align directly with business goals helps deliver immediate impact. As mentioned, the funding for this often comes from transforming operations, which is where we are actively supporting many of our customers.
CIO&Leader: NTT DATA invests around US $3 billion annually into R&D. How much of that is going into smart AI agents, and how do you measure return on these innovations?
Subhashini Desigan: As you mentioned, we invest more than US $3 billion annually in innovation.
One important aspect is that about three years ago, we began our globalisation journey. Earlier, we operated as more distributed organisations globally. Today, this globalisation effort allows us to consolidate investments more effectively and avoid duplication across regions.
For every investment, we have rigorous processes to measure ROI. This applies to smart agents, accelerators, and full-stack solutions.
One of NTT DATA’s strengths is that we take a long-term view of innovation. We do not evaluate innovation quarter by quarter. We focus on investments that shape industries while supporting sustainable, trustworthy, and ethical AI.
Some investments deliver immediate ROI, while others are designed for long-term transformation. We consciously maintain that balance.
CIO&Leader: How are you incorporating an AI-first strategy into infrastructure to better serve clients?
Subhashini Desigan: AI touches every function — offerings, sales, delivery, operations, finance, people, and culture.
Our AI-first strategy provides a unified view across all aspects of the organisation so AI adoption remains consistent, scalable, and embedded into the organisation’s DNA.
Customers and partners increasingly want proof that AI delivers sustainable business value.
Many organisations still keep AI buried within technical functions several layers down the organisation. Our approach is different because AI is driven from the top.
Scale requires urgency, direction, and top-down leadership.
Too many organizations stall at pilots and proofs of concept. With our CEO acting as Chief AI Officer, pilots can quickly scale into enterprise-grade offerings that deliver meaningful business impact.
Scale requires urgency, direction, and top-down leadership.
CIO&Leader: Could you tell me about the advantages your Centers of Excellence bring, and how the Indian ecosystem helps accelerate smart AI agent deployment?
Subhashini Desigan: Our Centers of Excellence focus on three key areas.
The first is training and enablement around specific technologies such as hyperscaler platforms, OpenAI, Mistral AI, and our own aXet platform.
The second is tools and accelerators. We have already built and deployed hundreds of smart AI agents on hyperscaler stacks, and we create reusable assets that accelerate deployment for clients.
The third is customer engagement and industry collaboration. We work closely with customers to understand where industries are heading and shape scalable solutions accordingly.
For example, on Salesforce, we launched Agentforce services and established a dedicated Center of Excellence while working closely with Salesforce on agentic transformation initiatives.
CIO&Leader: What operational bottlenecks have you identified while scaling AI-led competencies in India?
Subhashini Desigan: Investment prioritisation is an ongoing process. Decisions around resource allocation are constantly evaluated based on market demand, customer traction, and observed outcomes.
Training is also a critical part of the journey. Employees need dedicated time and environments to learn and engage with these technologies.
The enthusiasm among employees is very strong. The key challenge is creating the right conditions for learning while balancing ongoing responsibilities and directing investments into areas that deliver measurable business outcomes.
CIO&Leader: Looking ahead to 2027, what governance or operational challenges do you see enterprises facing while scaling AI-native architectures?
Subhashini Desigan: Foundational layers remain critical. Enterprises need strong digital cores, unified data strategies, semantic ontologies, and agentic layers where agents can operate effectively.
At the same time, the need for a single centralised data lake is becoming less important because agentic AI can work across multiple applications and both structured and unstructured datasets.
Investment prioritisation will continue to be a challenge for customers. Organisations need to identify areas where AI can generate immediate savings so those savings can fund future AI investments.