In the middle of chatter about AI replacing jobs in the tech sector, there is one company that is sounding a note of optimism for software engineers. This one is Zoho. When asked if AI can replace engineers at Zoho, the company CEO Mani Vembu says that the answer for now is No. That is because even though Vembu sees productivity gains coming from AI tools he doesn’t believe that AI can replace engineers and software developers at his company.

In a conversation with India Today Tech, Zoho CEO Mani Vembu clarified that job cuts due to AI are not on the horizon at Zoho because it is not yet capable enough. “If you ask the (AI) system to summarise a deal or generate content, it can do that pretty well. But replacing someone’s role entirely? I don’t think AI is ready for that yet,” he says.

And to emphasise his points, the Zoho CEO adds, “To be clear, we haven’t reduced any headcount due to AI. In fact, we are planning to hire more support engineers because our support volumes are increasing.”

Vembu made his comments at Zoholics in Bangalore recently. At the event Zoho unveiled its very own large language model (LLM) called Zia. This new AI model is pitched as India’s first enterprise-focused, home-grown AI system. While it’s not meant for the general public, the model is pitched to businesses and it comes integrated with Zoho’s productivity and enterprise platforms. Available in three parameter sizes (1.3B, 2.6B, and 7B), the LLM is designed to handle tasks like structured data extraction, summarisation, and prompt-based coding.

AI is becoming part of the workflow

Vembu believes that AI can be an enabler at work, helping employees do more. But it is not capable enough to do the entire A-Z of tasks. “So far (at Zoho) we definitely haven’t seen AI replacing employees,” he says. “What we are seeing instead is that it’s (helping) some of the roles. For instance, say a support rep typically handles 20 tickets a day, can we help them manage 25 instead? That’s a 20 per cent productivity increase.”

This is a point further explained by Ramprakash Ramamoorthy, director of AI at Zoho, in an exclusive conversation with India Today Tech. He says that Zoho not only uses the AI tools internally before releasing them to its customers, the company is also finding that productivity gains due to AI are real.

“We dogfood all the AI projects that we are about to launch,” says Ramamoorthy. “A lot of impactful areas are in summarisation or finding relevant information. For example at Zoho, a new recruit comes in and wants to find out which form to fill for a night care request. Previously, you’d have to ask somebody or your boss. Now, we have (internal) HRMS products with these bots that can guide them.”

According to Zoho, a similar impact is also being seen in customer-facing roles. “Even for a salesperson, aggregating information across the system is crucial. This is where we see the most impact,” says Ramamoorthy.

When asked if AI had in any way impacted Zoho’s hiring practices, Ramamoorthy categorically says no. “For Zoho, not yet,” he says, and highlights that the company continues to hire actively and sees AI as a tool for augmentation, not replacement.

AI challenges remain for Indian companies

While AI is helping Zoho to grow its business and make employees more productive, the company admits that Indian tech firms still face big challenges in building advanced AI systems such as Zia. Though impressive, in its capabilities Zia does not seem comparable to the likes of ChatGPT or Gemini.

Ramamoorthy shares that for Indian companies like Zoho there are mostly three major challenges in the AI space: compute, data, and expectation management.

“The first challenge is compute. It’s an expensive affair, and even if you put the dollars on the table, there’s a six-month lead time to it. And then there are import caps (on GPUs) for countries like India,” says Ramamoorthy. “The second is the availability of data. Consumer AI models work on consumer data, which is abundant. But in the enterprise, you can’t replicate that because businesses can’t just give away sensitive information. The third challenge is the over-hyping of technology. It’s not artificial, it’s not intelligence. Managing expectations is just as critical.”

– Ends