'GenAI giving 30-40% productivity gains'

For Dinesh Nirmal, who heads IBM Software, one of his mandates is to integrate generative artificial intelligence (GenAI) into all products IBM builds and leverage it to enhance developer productivity across global labs.

Nirmal believes that one of the biggest benefits of GenAI is automation, leading to optimisations and productivity gains.

He shares that IBM Software has observed productivity gains of 30-40 per cent in some segments of software development.

“Software development encompasses multiple aspects beyond code generation. In areas like code documentation, explanation, or generating test cases, we’re seeing 30-40 per cent optimisation and productivity gains,” he told Business Standard in a virtual meeting.

Nirmal says that IBM Software Labs is integrating GenAI into every product in IBM’s portfolio, empowering developers to write codes.

“Developers can now write hundreds and thousands of lines of code that they couldn't last year.

"It’s a tool to augment developer productivity, enabling faster code delivery and application development.

"For me, it’s about productivity and augmenting human capabilities rather than replacing them,” he said when asked about fears of GenAI replacing developers.

IBM Software Labs in India is located in Ahmedabad, Kochi, Bengaluru, Hyderabad, and Pune.

Nirmal illustrates how GenAI is being infused across IBM’s Indian labs.

“Our Kochi Labs is focused on automation products like watsonx Orchestrate. Ahmedabad is focused on integrating GenAI into our security products,” he added.

He also notes that a high percentage of developers at IBM’s Software India labs are proficient in GenAI.

“That’s why we’ve brought a lot more core missions to India.”

Despite the buzz surrounding GenAI, Nirmal says that its actual adoption and deployment have been limited.

“When we talk about GenAI adoption — not just proof of concepts but actual deployment in production environments — less than 10 per cent of enterprises have fully embraced it,” he said.

The reason is simple. He attributes this to enterprises having other concerns that take priority over adopting GenAI for end-user applications like ChatGPT.

“While the technology cycle is moving rapidly, deploying GenAI in enterprises will take time; that’s the reality.

"This is because a lot of the models also need to be vetted.

"If I go to Hugging Face, there are 6,000 models available.

"As an enterprise, how I choose any one is a complex issue,” he said.

IBM recently launched its GenAI Innovation Center in Kochi, aimed at enabling enterprises, startups, and partners to explore, experience, and build generative AI technology.

As organisations move from experimenting with AI to deploying it for business value, they often find AI projects too complex to integrate due to limited skills or expertise.

The GenAI Innovation Center will offer organisations access to IBM experts and technologies designed to help them build, scale, and accelerate the adoption of enterprise-grade AI.