More Code, More Review: How AI Is Changing Engineering

AI coding tools are generating code faster, shifting more of engineers’ attention towards designing systems, reviewing output, and testing whether software works reliably.

During a discussion titled ‘How Engineering Changes When AI Writes The Code’ at Inc42’s inaugural ‘The CTO Summit 2026’ in Bengaluru, panellists described how AI is changing engineering work beyond code generation, from the time spent on design and review to investment in testing infrastructure. This shift is also prompting teams to rethink productivity: more code does not necessarily mean better products or faster releases.

The panel discussion was moderated by Kshitij Shah, EIR at Digio.

Jitendra Agrawal, executive VP of engineering at JioHotstar, said engineers are spending more time on design and review, with greater emphasis on brainstorming before implementation.

He said AI allows teams to attempt features they had previously set aside but cautioned against claiming large productivity gains without evidence.

Pankaj Goel, CTO of LeadSquared, said AI can quickly inflate conventional measures such as pull requests per developer and lines of code without establishing whether teams are performing better.

“Some of these metrics get diluted with AI use and get inflated very, very fast,” he said. Speed to production, stability, availability, and customer impact, he added, are better indicators of engineering performance.

Testing Takes A Larger Role

At Cashfree Payments, spending on virtual environments for testing now exceeds spending on AI tools, according to CTO Ramkumar Venkatesan. The company keeps token spending constrained and ties it to returns, he said.

Cashfree has automated 97% of its test cases, he added, allowing its testing team to focus more on performance and security testing.

The emphasis on validation echoed a point raised by Razorpay’s SVP of engineering Prabhu Ram in another session at the summit: AI cannot make up for weak engineering foundations, including inadequate systems for monitoring performance and detecting problems.

Rethinking Access To AI Tools

Beyond measuring output and validating code, companies are also adjusting how employees access AI tools.

Thiyagaraj T, director of engineering at Eightfold, said the company moved from fixed AI budgets for employees to automatic top-ups after its initial approach resulted in high underutilisation.

“The common theme is that we want to democratise usage,” he said.

HR and operations teams now use higher plans than some engineers, he added, reflecting adoption beyond product and engineering.

The discussion highlighted why code generation alone offers an incomplete picture of AI’s value. Assessing the gains also requires accounting for design, review, testing, and the time it takes to deliver software that works reliably for customers.

The post More Code, More Review: How AI Is Changing Engineering appeared first on Inc42 Media.