AI Implementation As Challenging As Any New Technology: Nilekani

Every new technology wave comes with hype, but at an enterprise level, the difficulty levels will be similar or more like other technologies – Nandan Nilekani, veteran tech wizard

Bengaluru: Artificial Intelligence (AI)-led solutions will take time to be developed for solving specific problems of enterprises despite all the hype associated with the new technology.

Speaking at the Carnegie Global Technology Summit 2025 recently, Nilekani said that adopting AI is relatively easier in the consumer space, but at an enterprise level, the difficulty levels will be similar or more like other technologies.

He said that new technology and related hype usually go in tandem. “That was the case when cloud technology came in. When crypto came in, there were lot of talks about mitigating world hunger through the new technology. So, new technology and hype cycle are usual,” Nilekani said.

He, however, said that there are differences between hype related to AI. For the first time in human history, human beings are going to put their trust on machines for decision making. However, enterprise-grade solutions coming from AI will be difficult and take time.

companies are putting their brand behind an AI-powered offering and even minor errors can affect their image.

“Enterprise AI is much harder and it will take much time. The more difficult part is to implement AI in public sector. Because public sector has structural constraints. It has ministries, departments and everybody is territorial. So, data is not shared. Therefore, a lot of things need to be done if we want the AI stuff to work,” Nilekani said adding that adoption of AI solutions in the consumer space is going to be faster.

He noted that companies and governments are going to be careful before implementing AI solutions as they can’t afford wrong outcomes coming out the AI platform. “If the enterprise provides AI at scale and that AI even has 2 per cent error in the way it answers, then that affects the brand itself. That’s one of the reasons why it’s taking so long, because we don’t have the guardrails to ensure that we have absolutely no machine error in these things,” Nilekani said. According to the Infosys cofounder and architect behind India’s Digital Public Infrastructure (DPI), the challenges of implementation are the same like any other (technology).

“The hype is at a different level. AI doesn’t mean that it is easier to do. Rather, it is going to take the same effort or more effort. And as you are trusting the machine to make decision, there is more responsibility to see that it works,” he said. Nilekani, however, remained optimistic about India’s faster adoption of AI due to solid digital infrastructure.

"Indians should not worry about not having developed large AI models like some other countries. We shouldn’t be losing sleep because somebody has not built any AI models,” he said. According to the tech veteran, AI adoption can happen much faster in India because of higher acceptance level of Indians for new technology applications.

“Because of India’s situation today and technological sophistication that we are able to accomplish in the last 15 years or so, the gap between global development and AI adoption is going to be very short,” Nilekani said.

He also cited examples of Indian startup ecosystem, which has evolved over time leveraging technology.