AI shifts from experimentation to measurable business impact: NTT Data

Clients are seeking profit-and-loss gains from AI, using productivity savings to fund transformation as demand shifts from traditional developers to domain specialists

Artificial intelligence (AI) is moving beyond experimentation, with enterprise customers now demanding measurable business impact rather than proof-of-concept projects, according to senior executives at NTT Data. The company said clients are increasingly looking to use AI-driven productivity gains to fund broader business transformation, while demand is also shifting toward domain specialists who can work alongside AI rather than traditional software developers.

 

"The question from CEOs and boards has changed," said Amir Durrani, president (Global Application, Business Process Services and Data Practice) at NTT Data Inc. "Earlier it was about what companies were doing with AI. Today, the question is how AI is impacting the profit and loss statement."

 

According to Durrani, enterprises no longer have separate budgets for large-scale AI transformation. Instead, they are looking to generate efficiencies in their existing operations and reinvest those savings into business transformation.

 

He gave the example of NTT Data’s business process services segment. “In the claims management process of health clients, we are seeing productivity improvements of 60-70 per cent with AI assets, while application management and AI operations are delivering gains of 30-45 per cent. However, moving beyond those levels requires customers to allow the company to deploy its proprietary AI assets inside their environments,” he said.

 

India remains central to that strategy, with Durrani describing the country as the "heart" of the company's innovation and AI transformation efforts.

 

Avinash Joshi, executive managing director (India), said NTT DATA is already using AI to generate nearly five million lines of code in its cloud business for Indian clients and has rolled out enterprise-wide AI tools internally. The company has also committed to certifying 10,000 employees under the OpenAI partner programme and is empanelled under the India AI Mission to offer GPU-as-a-service.

 

Despite widespread concerns that AI-driven productivity will reduce hiring, Durrani said the company continues to add headcount, particularly in India. Instead of replacing employees, customers are redeploying experienced staff from routine operations into transformation programmes because of their deep understanding of business processes.

 

In India, the firms headcount is around 40,000. India is a key market for the Japanese firm.

 

The hiring strategy, however, is changing. NTT Data said demand is moving away from large numbers of Java or .NET developers toward domain experts who also understand AI. For example, one North American healthcare engagement requires radiologists rather than software developers, with the goal of turning them into Forward Deployed Engineers (FDEs). Durrani added that they already have thousands of employees performing FDE-type roles, although it is still working on formally defining the position.

 

On AI agents, Joshi said NTT DATA follows an "AI first, not human last" approach, arguing that people will remain in the loop as enterprises gradually adopt agentic AI. While organisations are increasingly measuring success by the number of AI agents being developed and deployed, he said governance, ethics and hallucination management remain critical before human involvement can be reduced significantly.

 

Durrani added that enterprises are also becoming increasingly sensitive to the cost of AI itself. Beyond productivity, customers are now seeking ways to optimise token consumption across different AI models. NTT DATA is positioning capabilities such as its Tsuzumi model and AIVista to fine-tune models on customer data while reducing token costs.

 

The company also sees enterprise technology engagements gradually moving toward business-outcome-based pricing, although executives acknowledged that most customers are still reluctant to hand over enough operational control for such contracts. NTT DATA already operates outcome-based engagements in areas such as contact centres and claims processing, where payments are linked to successful outcomes rather than traditional effort-based metrics.

 

Looking ahead, Durrani said AI is putting pressure on traditional IT services revenue, but the company expects that to be more than offset by demand for AI-led transformation projects. "Overall, we believe we are on a growth trajectory, and it is mainly AI-led transformative business," he said, adding that growth, rather than headcount, is becoming the more meaningful measure of an IT services company.