A 39-year-old accountant knew AI was coming for his job — so he started vibe coding

Wei Khjan Chan, a seasoned accountant with 18 years of experience, has taken proactive steps to adapt to the looming threat of automation in his profession. Facing a future where AI could potentially replace jobs like his, Chan decided to take matters into his own hands by exploring the concept of vibe coding. After attending coding workshops in Singapore and Malaysia, Chan, who lacks a formal technical background, discovered vibe coding in June. He quickly realized its potential to streamline his work processes. He developed a web application designed to simplify the tedious task of filing expense claims after business trips. This innovative app utilizes AI-driven optical character recognition, enabling it to scan and process receipts, which are then automatically organized into files for his company's finance department. In addition to streamlining expense claims, Chan is leveraging AI to enhance his overall workflow, including automating invoice generation. "This code is a bunch of JavaScript, which obviously I don't understand," he remarked while demonstrating his web app. He emphasized that without the tools and skills learned through vibe coding, accountants would struggle to achieve such efficiencies. Unlike others who might view vibe coding as a means to pivot careers, Chan believes that understanding AI is becoming a vital skill for all office professionals, much like proficiency in Excel. He noted that the ability to create apps has transformed the way projects are developed; tasks that once took weeks can now be executed in just a weekend. As a committee member of his local accounting institute in Malaysia, Chan is an advocate for increased AI training within the industry. He has observed a decline in the number of individuals entering the accounting field, despite a growing demand for services. He believes that AI could serve as a crucial resource to bridge this manpower gap. Chan shared insights from his early experiences with AI, noting that while he was initially advised to provide long, detailed prompts, he found that smaller, incremental adjustments yield better results. He likened the process to managing an intern, where breaking down tasks into specific instructions leads to improved outcomes. However, his journey hasn't been without challenges. He once faced a significant setback when he built a database tailored for a single organization, only to realize later that it needed to support multiple companies. This experience underscored the importance of establishing a robust architecture from the outset. Chan's approach to debugging also reflects his adaptive mindset. He humorously described it as "complaining to the AI," where changes in error messages indicate progress, while persistent issues prompt him to reset and reframe his requests. Despite the challenges, Chan finds joy in vibe coding, often working on his projects after his children have gone to bed. He compares the experience to playing a game, where incremental improvements lead to a coherent final product over time. Through his journey, Chan exemplifies how embracing AI can redefine traditional roles and enhance productivity in the workplace.