Financial services firm Money Forward has significantly expanded its use of Cursor coding agents, moving beyond its engineering department to integrate the AI tools into product, design, and quality assurance workflows. This strategic rollout aims to embed AI assistance across the entire software development lifecycle.
Initially adopted by engineering, Cursor agents are now saving developers an estimated 15-20 hours per week on tasks ranging from refactoring to cloud deployment management. This success spurred a company-wide evaluation, leading to broader adoption across over 1,000 employees.
From Engineering to Everything
Money Forward’s Engineering Productivity and AI Research (MEPAR) department evaluated multiple AI coding tools. They selected Cursor for its model-agnostic infrastructure, which allows asynchronous cloud agents to connect to internal tools for rapid context retrieval without local hardware limitations. This flexibility proved crucial for teams beyond engineering.
Key advantages cited include minimal setup, enabling immediate use across departments with varying technical expertise. Cursor's built-in browser was also instrumental, allowing designers and QA engineers to visually verify changes, a feature lacking in terminal-based alternatives.
The unified agent workspace, handling code generation, review, testing, and debugging in one place, eliminated tool-switching friction. Crucially, Cursor's ability to perform reliably on Money Forward's complex, interconnected production codebases facilitated adoption by non-engineering teams interacting with code for the first time.
Streamlining QA and Product Development
QA engineers have seen a 70% reduction in test case generation time. Previously manual processes are now automated, with agents generating structured test cases from Jira tickets and Notion documents, and then translating them into Playwright scripts. This shift allows QA to focus earlier on risk-based testing and quality gates.
Product managers leverage Cursor to analyze production code, extract system relationships, and generate architecture diagrams. This grounds requirements in implementation details, helping identify edge cases before development begins and improving overall SDLC efficiency. Even without existing documentation, agents can infer specifications directly from code.
Design Iterates Directly on Live Code
Designers are now working directly against live frontends, using Cursor's full-stack context and browser capabilities. This contrasts with historical workflows relying on static mockups and secondhand descriptions. Designers also access product analytics through Cursor to refine designs based on real user data and product behavior.
This expansion highlights how Cursor coding agents are proving valuable beyond core development, impacting broader product and quality functions. The firm's move mirrors trends in AI in financial services, where advanced models like those powering Cursor's capabilities, alongside options such as ChatGPT, Claude, and Gemini, are reshaping workflows.
