Two weeks ago at Risk.net Asia Risk Congress, I spoke about AI adoption in the financial industry, and particularly the complexity to move from AI experimentation to enterprise transformation.
When ChatGPT launched nearly 2 years ago, the excitement was palpable. Even heavily regulated and conservative financial institutions jumped at the opportunity to explore AI, particularly GenerativeAI.
Now that the dust is starting to settle, here’s what I’m seeing as a fractional CDO:
โ
Successful POCs like AI-driven chatbots, automated document parsing (for example to enhance Know Your Customer (KYC) processes)
โ Stalled progress in scaling these successes enterprise-wide
๐ค Leadership wondering: “Why isn’t this working at scale?”
๐ช๐ต๐ฎ๐’๐ ๐บ๐ ๐๐ฎ๐ธ๐ฒ?
The roadblocks aren’t in the AI technologyโthey’re in the foundation:
๐งญ Strategic Misalignment: Too many institutions treat AI as a tech project, not a business transformation
๐๏ธ Legacy Architecture: It’s hard to build a superhighway on dirt roads
๐ Data Silos: You can’t fuel AI with data trapped in organizational vaults
๐ง๐ต๐ฒ ๐ฃ๐ฎ๐๐ต ๐๐ผ๐ฟ๐๐ฎ๐ฟ๐ฑ
1๏ธโฃ Data Readiness:
For AI to work at scale, a strong data foundation is critical. Financial institutions need to break down silos, modernize data infrastructures, and invest in clean, accessible data.
2๏ธโฃ Cultural Shift:
Beyond technology, leadership must drive a shift in culture to treat data as the true strategic asset it is.
3๏ธโฃ Regulatory symbiosis:
In Singapore, we’re lucky that regulatory bodies like MAS are working closely with financial institutions to build trust in AI. Initiatives like FEAT principles (Fairness, Ethics, Accountability, Transparency) and Project Veritas are paving the way for more responsible AI adoption.
๐ง๐ต๐ฒ ๐ข๐ฝ๐ฝ๐ผ๐ฟ๐๐๐ป๐ถ๐๐
Financial institutions that get this right aren’t just adopting AIโthey will have the opportunity to reimagine their entire operating model.
As a Fractional CDO, I help companies bridge the gap between AI ambition and enterprise-wide execution, by focusing on Data first.
