PineLabs Reports 89% AI Agent Contribution to Code

Pine Labs is leveraging its partnership with OpenAI to build an "Agentic Commerce" ecosystem, moving beyond simple automation to create a reasoning layer for financial transactions. This allows AI agents to autonomously handle complex tasks like negotiating supplier terms and optimizing cross-border settlements within secure financial guardrails. The company's CEO, Amrish Rau, has stated that due to these AI-driven productivity gains, Pine Labs has frozen its engineering headcount at 1,000. As of late 2025, Rau noted that 18% of all code at the company was already being written by generative AI, with some legacy platform rewrites seeing 40% AI code contribution. The integration of custom AI agents across the development lifecycle is key to this strategy. For instance, agents can convert Figma designs directly into front-end code, streamlining the handoff from design to development. This approach treats the design file as the single source of truth, reducing manual work and potential errors. In the coding and review phases, AI agents are integrated with GitHub to automate pull request summaries, conduct initial code reviews for bugs and vulnerabilities, and ensure adherence to style guides. This frees up senior engineers from routine checks to focus on more complex architectural decisions. For project management and testing, AI agents in Jira can automate ticket creation and triage, generate user stories, and even create test cases directly from requirements documents. In testing, AI can intelligently prioritize test execution based on code changes and historical data, significantly speeding up the QA cycle. This shift towards AI-centric development fundamentally alters how engineering efficiency is measured. Traditional DORA metrics are being re-evaluated as AI can dramatically increase deployment frequency and reduce lead times, pushing leaders to focus more on the business impact and quality of AI-assisted output.