Business News/ Companies / Building right organizational strategy a big challenge in AI adoption: TCS CTO
Gift this article
Building the right organizational strategy by getting all teams together and modifying the role of human employees is the biggest roadblock to adoption of artificial intelligence by businesses, Harrick Vin, global chief technology officer at Tata Consultancy Services (TCS), said in an exclusive interview with Mint.
Building the right organizational strategy by getting all teams together and modifying the role of human employees is the biggest roadblock to adoption of artificial intelligence by businesses, Harrick Vin, global chief technology officer at Tata Consultancy Services (TCS), said in an exclusive interview with Mint.
Speaking about the core challenges in the enterprise adoption of AI and generative AI technologies, Vin said, “Organizations must design their AI strategy to balance value, complexity and risk. To do this, they need to get multiple teams together, which makes it a big challenge for the company itself, since each team will have different requisites and philosophies. But, for a cohesive AI strategy to become AI-mature, creating a common take is crucial," he said.
Speaking about the core challenges in the enterprise adoption of AI and generative AI technologies, Vin said, “Organizations must design their AI strategy to balance value, complexity and risk. To do this, they need to get multiple teams together, which makes it a big challenge for the company itself, since each team will have different requisites and philosophies. But, for a cohesive AI strategy to become AI-mature, creating a common take is crucial," he said.
Also Read: Lenders get serious on AI with in-house talent, outside help
“This is why the adoption of AI is not up to expectations. A relatively small percentage of use cases are actually going into production, because they’re failing on not being engineered to a company’s needs, or a company not knowing how to drive the organizational changes to take AI deployment from experiment to value-driven production."
To help improve this, Vin said that while TCS, India’s largest software services company by market cap and revenue, continued to incrementally raise its research and development (R&D) expenditure, there is unlikely to be any exponential jump in this figure. In FY24, TCS spent ₹2,751 crore on its R&D initiatives—about 1.2% of its ₹2.41 trillion annual net revenue.
“We’ve stayed around the 1.2% figure for a while, keeping our research expenditure consistent. It is difficult to project any change for the future—right now, the magnitude of our investments is not limited by any arbitrary figure, but based on where we need to invest in R&D. We’re constantly observing various industries, which helps us decide upon where and when to invest," Vin said.
The affirmation of TCS’ R&D approach to address technical challenges comes against the backdrop of India looking to set up a target-linked incentives programme to help private sector companies increase R&D spending. The programme, senior government officials told Mint last week, is designed to help create technical patents and intellectual properties within India—instead of licensing tech from overseas economies. In a proposal for a $3-billion incentives package to boost India’s tech industry, up to $1 billion could be earmarked to boost private-sector technical R&D expenditure, Mint reported on 12 December.
Focus on R&D to address AI challenges
Vin, meanwhile, said that TCS is using its R&D expenditure to develop solutions that make AI more palatable to enterprises.
“Our R&D approach is divided between the future of tech initiatives, and also the future of work. For the latter, there is knowledge work, and physical work. There’s also the future of societal work that defines areas such as food, agricultural development, waste management and more. In physical work, we split our research into areas including the future of robots and robotic automation, and automation of factory operations," he said.
Also Read: Generative AI, data centres to define India’s tech industries in 2025
He also identified six key fields that defined technical innovation demand from enterprises through this year. “We have sensing technologies to collect and observe lots of data globally—without data, there’s no machine intelligence. The second is the rise of quantum and neuromorphic computing that brings brain-like computing power to the edge at nearly a hundredth the power consumption—giving rise to us working towards a future where all computing will be hybrid. Data management and security is an obvious third, given all the threats. In communications, 6G technologies will set standards; in AI, knowledge management to upskill the future of our work is happening as we speak. Finally, new human-computer interfaces will become key as we work even more closely with machines," Vin added.
As an example of how such initiatives are playing out, Vin highlighted the ongoing project examples of using hyperspectral imaging, that is, data from satellites, to assist in city planning projects. Further, the presence of low-power processors and local AI models help connected cars gauge parameters to detect if a driver is fatigued.
All of this, Vin added, will depend on the evolution of AI models, and the human understanding of deploying the right models for the right use cases.
“Traditional AI models are inherently more predictable, explainable and efficient, since they are based on causality and logical reasoning, and are deterministic. Generative AI models, on the other hand, use analogical reasoning and correlational learning. The advantage of the latter is that it is inherently more adaptive than rule-based models—on the other hand, they’re non-deterministic and not explainable. A key challenge is to know when not to use generative AI, and also the right model to use. Agents, on this note, help determine the outcome of an ensemble of models, tally that with external environments, and drive outcomes. These agents are designed for change, failing which they can become obsolete quickly. From an engineering standpoint, this is why AI is a big challenge—and thus akin to an art form today," he said.
Generative AI, despite its hype, is yet to create a major revenue impact for tech exporters. On 19 December, Accenture Plc reported $1.2 billion worth of generative AI bookings or orders. It has recorded $500 million in generative AI revenue in the latest quarter.
Also Read: Smartphone market stays below pandemic levels, concerns remain
While TCS disclosed 600 active generative AI projects in its September-quarter earnings press conference, a quarterly revenue figure wasn't shared.
With the current initiatives, TCS’ AI revenue could swing upward—even in the absence of accelerated R&D expenditure from the domestic tech bellwether.
Catch all the Business News , Corporate news , Breaking News Events and Latest News Updates on Live Mint. Download The Mint News App to get Daily Market Updates.
Get the latest financial, economic and market news, instantly.
{{^adFree}}
{{/adFree}}