‘90% Of The Code We Ship Has No Human Touch’: Emergent On AI-Powered Software Development
Artificial intelligence is increasingly moving beyond helping developers write code to taking on a much larger part of the software development process. According to Mukund Jha, Founder of Emergent, around 90 per cent of
Ni Debashis Sarkar··34K nabasa
‘90% Of The Code We Ship Has No Human Touch’: Emergent On AI-Powered Software Development
Artificial intelligence is increasingly moving beyond helping developers write code to taking on a much larger part of the software development process. According to Mukund Jha, Founder of Emergent, around 90 per cent of the code his company ships now has no direct human touch.
Jha made the remarks during a panel titled “India’s AI Unicorns: Models, Moats & the Race to Scale” at the Moneycontrol Startup Conclave 2026 on September 26. The discussion was moderated by Nalin Mehta, Managing Editor, Moneycontrol, and also featured Pratyush Kumar, Co-Founder of Sarvam.
‘90% Of The Code We Ship Has No Human Touch’
Speaking about how AI is changing software development inside Emergent, founder Mukund Jha said that the company is already seeing a significant shift in the amount of work that can be handled by AI systems.
“90% of the code we ship has no human touch,” Jha said during the discussion.
He explained that AI is not simply being used as another coding assistant. Instead, the technology is increasingly involved in producing working software, changing the way developers and companies approach the development cycle.
Jha gave an example of how quickly this can happen. A feature that the company initially estimated would take around two to three weeks was completed in two days, according to him.
The example highlights the speed that AI-powered software development can bring to teams, particularly when AI systems are able to take on more of the actual coding work rather than simply suggesting individual lines or snippets of code.
AI Is Changing The Role Of Developers
Jha’s comments point to a broader change in the way software could be built as AI models become better at reasoning and coding.
During the panel, he spoke about Emergent’s development from software creation towards what he described as more autonomous business automation. The company is building AI agents that can carry out tasks rather than simply responding to individual prompts.
According to Jha, more than 30,000 businesses use one of Emergent’s persistent agent products every day. He also said the company is seeing faster growth in Western markets, while India currently contributes around 8-10 per cent of its revenue.
The growing capabilities of AI coding systems also mean that the traditional boundaries between writing code, testing software and deploying features are beginning to change. Instead of developers manually handling every stage, AI can increasingly participate in much of the process.
The Risk Of ‘AI Slop’
However, Jha also pointed to a downside that comes with the ability to generate software at much higher speed: the possibility of producing large amounts of low-quality or unnecessary code.
He referred to this as “AI slop”, highlighting the concern that faster software development does not automatically mean better software.
If AI can generate code significantly faster than traditional development teams, companies may be able to build and release features more quickly. But the quality of those outputs, along with security and reliability, remains an important consideration as AI takes on more of the development process.
Open-Source AI Is Also Improving
Jha also discussed the progress being made by open-source AI models. He said improvements in reasoning and coding capabilities are making these models increasingly useful for software development while also helping improve margins.
This is important for companies building AI-powered products because access to capable models is no longer limited to the largest technology companies. As models improve and become more accessible, startups can use them as part of their own products and workflows.
For Emergent, that is part of a wider shift from AI-assisted software creation towards systems that can take on increasingly autonomous tasks.
India’s AI Race Is Also About Cost And Scale
The discussion also touched on the wider opportunity for Indian AI companies. Sarvam’s Pratyush Kumar said India is effectively entering what he described as “year one” of its AI journey, with the country now building capabilities around models, GPUs and talent.
Kumar also stressed the importance of reducing the cost of AI and building applications that can work at India’s scale. He cited potential use cases across areas such as agriculture, education and healthcare, where AI could potentially provide access to intelligence to very large numbers of people.
For companies such as Emergent, the software-development side of this transformation is already happening. Jha’s “90%” comment offers a glimpse into how quickly AI is becoming part of the actual process of building products, rather than remaining a tool used only by individual developers.
The bigger question is how far that shift can go as AI models become more capable, and how companies balance the speed of AI-generated software with the need for quality, security and human oversight.