Across major tech players, AI is writing large swaths of code, with India’s Meesho reporting well over 70% of code AI-generated and Freshworks citing more than half, signaling a shift from experimentation to routine workflow.

AI is moving from experimental use to everyday development, increasingly handling drafts, tests, reviews, deployment fixes, and monitoring, while humans review for correctness and security.

In practice, AI mainly produces first drafts and routine code; engineers focus on review, design, architecture, and solving complex problems, leading to smaller, faster teams.

Anthropic says Claude is the primary author of development, with engineers pushing AI-generated pull requests in the low-thousands of lines and near-total AI code authorship anticipated per leadership discussions.

Banksathi reports about 80% of work done by AI, suggesting digitization will focus on standard processes going forward.

The overarching takeaway is that AI-enabled coding boosts productivity and alters team dynamics, but governance is needed to manage quality, risk, and accountability.

Meta targets show broad AI-assisted coding, aiming for 65% of engineers to write over 75% of their code with AI by early next year, with internal teams forecasting 50–80% AI-assisted coding.

Microsoft notes AI writing roughly 20–30% of code, while leaders project AI could handle up to 95% within five years, reserving human effort for creative and problem-solving tasks.

Experts caution that rising AI-generated code brings challenges: potential job shifts, bugs, security concerns, accountability gaps, and organizational changes.

Google reports AI generates about three-quarters of new code, with engineers performing subsequent reviews, marking a sharp rise from past years.

Fintech firms Credgenics and Paisabazaar highlight AI handling majority of backend work, credit workflows, and even customer interactions, including AI-generated videos and chat services.

Across the sector, AI-driven coding is accelerating productivity and reshaping teams, underscoring the need for governance and alignment with product goals.