K Krithivasan, who is set to complete two years as the Chief Executive Officer and Managing Director of Tata Consultancy Services in March, has had a tough period with global tech spends slowing down. However, he is optimistic that 2025 will be a better year with AI beginning to move the needle as clients think about technology modernisation. businessline spoke with Krithivasan about why he is optimistic and how AI deployment is taking shape for TCS and other companies worldwide.

Edited excerpts:

You’ve interacted with a lot of CEOs in India and globally as well. Do you see them now talking about reviving spend on tech?

There are two or three things that have been holding them back. Obviously, the cloud transformation has to happen. If they have to really leverage AI, the data estate has to be much better. Data should be available and accessible. You have to have the ability to use your data with external data, etc, and then come up with the proper recommendations. In addition, many organisations know there’s a high technology debt, but we are hearing more discussions about technology modernisation.

In 2023, you had said that AI is still not changing the needle for you and the industry. Has this changed?

The needle is beginning to move. There are more programs going into production and the maturity of adoption has improved. Understanding what AI can do for them is also evolving. If you try to use AI for a very small use-case across a wide value chain, it’s not going to change anything because you cannot have super efficiency in only one small process, whereas everything else is inefficient. So, we need to ensure that you are consuming in the right way and that there is efficiency all across.

The next important thing is, there are areas where new products and new services could be launched, something that could not have been done without AI. For instance, if somebody is selling insurance, generative applications that can actually do 100 per cent compliance check for all voice calls and give feedback on training. I think such use-case will continue to evolve and become more mainstream. After some time,  you may not want to call it an AI application since every application will have some component of AI infused in it. So, there will be elements of AI both in the build process as well as in the final outcome.

With deals, do you measure how much of that is coming from AI? Are you quantifying it?

We do track internally the engagements we are doing, but it is more to see whether are we doing it across all industries and clients. We believe that we will soon stop tracking the revenue coming exclusively from AI. If I sign a $100 million application maintenance deal, it will have a component of AI. It will become a very theoretical exercise to say that in this $100 million, should I attribute $10 million to AI? But we need to ensure anything we do leverages the AI subsystems.

So internally how are you using AI, like in terms of your processes and systems?

I am quite excited about AI in processes, where somebody undergoes a training on a particular subject. Some of these AI agents are able to tell you not just how you articulated but how you should have articulated. We have traditionally been using resume checkers and other enterprise applications. There is a reasonably strong internal push for AI. It has not changed the way we are hiring but it may after some time.

There’s an on-going debate among global leaders on whether Indian companies should be at the forefront of AI innovation or just adopt and build on what already exists. As a leader of IT tech in India, how do you see this approach towards India stack?

I don’t think there’s going to be a huge incremental advantage in building your own LLMs since there are so many already available. If you go by the existing heritage of most Indian organisations, we are more of system integrators. We use the products as a software and ensure that the benefits are given to the customer. If you take a similar approach here, we are better off—at least in the short run—using the LLMs available, but ensure that enterprises are able to use them to really get the value. There are opportunities to develop our own LLMs when regionalising it or building Indian language models, ensuring that Indian knowledge base is also captured in those models. Once you do that, you can bring in more inclusion and democratisation. So, we have to leverage for customers what is already available and create new markets. Doing it for Indian languages can create new markets in India and deliver services that could not be offered to the masses.

You had done a partnership with Nvidia itself for developing some of these AI solutions and training. So, how is that progressing?

It’s progressing quite well. We have chosen few industry segments. Currently, we started banking, financial services and manufacturing and we are developing solutions on those in India stack.

What is your take about India’s AI regulation?

My view is that every country has to have them because there is always a risk of misuse. So, we do need to ensure the responsible usage of AI. Since these are all early stages you won’t know what is the right amount of regulation, we’ll soon learn. We need a regulation in terms of protecting the intellectual property.

On the BSNL deal, can you talk about how it actually shaped up and whether you are expecting more 4G deals, especially from private players?

The deal has progressed quite well, with close to 70,000 sites installed, and about 60,000 carrying commercial traffic. The last set of sites will be more difficult as they will be in more remote areas. But we are on track to complete them.

The reason that we got into the deal is based on the capability we built to offer this solution to other private players, within and outside India, both on the 4G and 5G stack. Because of the unique and indigenous nature of the solution, it makes the country independent from external influences on critical infrastructure—which, we think is a very strong value proposition.

You will complete two years in March. What is your experience from your biggest challenges or learnings you’ve had in these two years?

Yes, I’ve been in this role for two years but I’ve been with the company for 34 years. As CEO, majority of the people I work with are in the US and I do interact with more external stakeholders. It’s been a great experience with lots of learning, in terms of understanding expectations from different stakeholders and how the society looks up to TCS. In fact, there is an expectation in terms of being a part of the Tata Group. I have not achieved all that I set out to achieve because, from a growth perspective, you want to have the biggest growth year-on-year. So, we have made good progress on the margin front. On the top line front, we have not achieved the double-digit growth. Although the market conditions have been challenging, we are aspiring for it.