A piece of technology won’t replace you overnight: Zerodha CTO on AI Policy and job security

Kailash Nadh, CTO, Zerodha

Zerodha Chief Technology Officer (CTO) Kailash Nadh took a firm stance on the role of Artificial Intelligence (AI) at workplace, emphasising that automation should enhance jobs, and not erase them.

Speaking on Zerodha's AI policy introduced years ago, he explained how it was designed to reassure employees that technological advancements wouldn’t jeopardise their roles.

"Two to three years ago, when the big LLM breakthroughs happened, there was a lot of fear, uncertainty, and hype. Employees were afraid of joblosses. So, we came up with an AI policy that says the introduction of AI or any automation will not directly lead to joblosses in the team, and there will always be opportunities. There will always be new roles," Nadh recounted at Moneycontrol Global AI Conclave in Bengaluru. “We wanted to pacify and assure our people that a piece of technology wouldn’t come and displace them overnight.”

Nadh further shared that Zerodha hasn’t really found any use cases for deploying some of these technologies, especially Gen AI models, in consumer-facing products or features. However, from an organisational efficiency perspective, the company has made massive strides with simple workflows built out of open-source models.

One of Zerodha’s notable AI implementations is in customer support quality assurance. The company leveraged open-source AI models to automate the transcription and analysis of customer calls, a task previously managed by a 100-member team. Nadh explained how this automation enhanced efficiency and reduced monotony.

“We’ve cobbled together an open-source system where recorded calls are transcribed using locally hosted models, voice becomes text, and then LLMs analyse the text. The system flags calls that need attention based on quality parameters, so the team doesn’t have to randomly sample and listen to thousands of calls anymore. This has made their lives better, allowing them to focus on ensuring quality.”

Asked if automation made the 100-member team redundant, Nadh clarified, “Those roles technically are redundant, but they’ve transformed into doing quality assurance now. Honestly, we don’t really need a 100-member team, but our AI policy ensures that technology doesn’t displace jobs. Instead, it creates opportunities for new roles.”

Nadh also spoke about the risks of indiscriminate automation in areas of critical decision-making. “Imagine an insurance claim being denied by a black-box AI system that can’t explain its decision. Over time, people working around such systems lose agency. Accountability and explainability are critical when AI tools are used for life-and-death or financial decisions,” he said, adding that Zerodha ensures that humans remain accountable in such scenarios.

Nadh was joined by fintech major PhonePe's CTO Rahul Chari. On the impact of AI on jobs, Chari expressed a balanced perspective, highlighting that while AI automation can lead to job displacement in certain areas, it also creates opportunities for higher-value roles.

He explained that AI is being used at PhonePe to enhance efficiency in tasks like customer support and merchant onboarding, but not as a replacement for human intervention in critical areas. “AI is there to assist and augment human capabilities, not to completely eliminate them,” Chari said, underscoring the importance of retraining and upskilling the workforce to adapt to the evolving landscape.

(L-R) Rahul Chari, CTO at PhonePe, Kailash Nadh, CTO at Zerodha along with Chandra R Srikanth, Deputy Executive Editor, Moneycontrol

AI in recruitment and development 

Despite the AI revolution, Nadh said Zerodha’s approach to hiring remains unchanged. “Most code, dev, and tech-related AI models have become commodities. If you have smart and curious developers, they’ll use these tools to improve their workflows. Nothing’s fundamentally changed in how we hire engineers,” he noted.

However, the CTO acknowledged that AI has simplified boilerplate programming tasks, allowing engineers to focus on more complex, innovative work. “Certain kinds of programming work, like boilerplate tasks, can now be automated. Entry-level work is being reduced, but models still can’t create well-engineered, future-proof systems. They’re tools for efficiency, not replacements for human ingenuity.”