Zoho co-founder and Chief Scientist Sridhar Vembu has shared an example of AI’s productivity revolution in software development. In a detailed post on X (formerly Twitter), the famed co-founder, who worked with his team to create the Arattai messenger app, has now revealed how one experienced engineer at the company independently created a sophisticated security tool in just one month – a task that would traditionally require a team of three to four engineers nearly a year to complete.
Vembu expressed his astonishment at the achievement, noting that he was completely unaware of the project until the engineer demonstrated the finished tool. “I was blown away by the depth and breadth of the tool,” he wrote. Vembu says that the engineer developed the complex assembly and machine code security tool in his spare time over almost four weeks, without any formal assignment or structured oversight.
Zoho engineer uses Claude Opus 4.5 for creating new tool
According to his post, the breakthrough in coding was powered by Anthropic’s Claude Opus 4.5 AI model, which the engineer described as a true “game changer.” Previously skeptical of AI-generated code, the developer found that Opus 4.5 dramatically accelerated experimentation, iteration, and refinement, compressing what was once a protracted development cycle into weeks.
Vembu compared the shift to the industrial revolution’s impact on textiles. “This is the arrival of powerful machine looms for software development,” he said, comparing AI-assisted coding to mechanised looms that disrupted traditional handloom weaving. He warned that this transformation challenges long-standing assumptions about team sizes, timelines, and productivity in software engineering.
“At Zoho, we have our work cut out, and as Chief Scientist, I am under pressure. Wish us luck,” he wrote at the end.
Vembu fostering a culture of AI innovation at Zoho
Previously, Vembu has stated that Zoho is rolling out AI coding tools across its engineering teams, but with a cautious philosophy for engineers – no mandatory usage. He also said that all AI-generated code must be thoroughly reviewed and approved by humans who take full responsibility.
He has shared that engineers at Zoho were coming up with productivity gains after relying on AI tools. Veteran engineers were completing performance-critical UI tasks in a single day that once took three weeks, and junior developers were building impactful internal tools they couldn’t have created without AI assistance.
Vembu had also stressed that companies must balance human expertise with growingreliance on Ai and machine learning to stay competitive in this new era.
