The Taiwan Banker

Whenever an interviewer asks NVIDIA CEO Jensen Huang how he approaches complex challenges, he often responds with the mantra of “first principles thinking.” NVIDIA’s success today is closely linked to this mindset, which has also been popularized by business leaders like Tesla CEO Elon Musk. Put simply, this means continuously asking “why” until one reaches the fundamental essence of an issue, avoiding relying on the momentum of the status quo. Historically, the earliest proponent of this approach was Aristotle, the Greek philosopher renowned for his questioning spirit. In Metaphysics, Aristotle argued that to truly understand something means returning to the most foundational truths: principles that cannot be further reduced. He believed we should not rely solely on analogies and rules of thumb (such as “fire = heat = rises”), but instead rebuild logic and knowledge from basic principles, and then reason upward to generate new understanding and solutions. In the years since AI has entered the mainstream, organizations have moved from early surprise and skepticism to initial adoption, and now to a new reality: the question is no longer whether to adopt AI, but how. How much should be invested? With limited budgets and unclear short-term profitability, how should priorities be set? Particularly in the highly regulated financial sector, AI adoption is not merely about improving operational efficiency, but also requires careful consideration of risk management, compliance, and anti-money laundering (AML). This is precisely where first principles thinking becomes valuable. Decision makers face an unprecedented transformation. The ability to reason from first principles may determine an organization’s future. This issue of the Taiwan Banker aims to help readers cut through this fog. While each contributor presents their analysis from a different professional perspective, their arguments converge on one shared conclusion: AI decisions can no longer be made through linear thinking, incremental iteration, or siloed implementation – the approaches which shaped the previous generation of digital transformation. This wave of innovation will be about structural innovation and rebuilding of value chains. Banks must dismantle complex workflows, step by step, continuously asking why a customer needs a given financial service in the first place. In the previous transformation, banks largely migrated branch services onto mobile apps, a type of process innovation which did not necessarily challenge the essence of finance itself. This time, however, AI invites a deeper question, prompting both creation and disruption: what is the purpose of finance? Without breaking operations down to their irreducible core, instead of transforming an organization, AI may end up simply exhausting employees. Ultimately, success depends on whether decision-makers are willing to reset existing business models and processes, to “zero out” legacy assumptions, and to resist the temptation to rely on past formulas for success, such as fragmented or agile project delivery in isolation. Rather, they must commit to a systematic operating architecture. Win Way Ngo, Assistant Research Fellow at TABF and the lead author of this issue, emphasizes that the true differentiator is not how impressive a single project appears, but whether a bank has the determination to invest in a robust AI platform, connecting scattered proof-of-concept (POC) initiatives into a shared foundation of scalable capabilities that generate synergy across the organization. Otherwise, AI projects will remain in a perpetual stage of experimental demonstration, ultimately leaving organizations vulnerable to competitors who have deeply integrated its new capabilities.