Artificial intelligence is entering a new phase in financial services as institutions shift from isolated experimentation to broader organisational integration. This transition is forcing leaders to confront deeper questions about capability, governance and the long-term structure of their businesses. TAB Global and Huawei convened a closed-door discussion aboard EagleWings V, departing from ONE°15 Marina at Sentosa Cove and sailing toward the Southern Islands. The vessel paused in the calm bay between Lazarus Island and St John’s Island where the discussion continued. As it later moved towards the industrial outline near Pulau Bukom, participants reflected on how AI is likely to redefine the foundations of financial services in the decade ahead. Huawei’s Jason Cao, chief executive officer of the Digital Finance Business Unit, shared several insights that captured the direction institutions are taking as AI becomes a core capability. AI shifts from narrow implementation to structural change Cao observed that many institutions are moving beyond viewing AI as a collection of tools. They are starting to recognise that the technology alters how information flows, how work is organised and how decisions are made. This reflects a shift from project-based adoption to structural transformation. He noted that early efficiency gains are only a small part of the long-term impact. Institutions now face questions about how to redesign processes and how intelligence will operate across front, middle and back-office functions. These changes require leadership that is willing to rethink traditional operating assumptions. Cao added that institutions are beginning to treat AI as an operating layer that will support or guide entire workflows. This implies new responsibilities for human oversight, particularly around governance, risk management and interpretation. Participants discussed how this shift may influence customer interaction models, risk processes and the distribution of authority between machine-driven and human-driven judgement. Cao reflected that this marks the point where institutions must think about enterprise-wide coherence instead of isolated optimisation. Democratising AI while strengthening governance becomes essential A recurring theme in the dialogue was the need to make AI accessible across the enterprise while maintaining clear oversight. Participants noted that institutions will not scale AI effectively if ownership remains concentrated within technical teams. One participant highlighted the importance of enabling business units to use AI independently, supported by strong governance, standard frameworks and defined guardrails. This approach gives teams the ability to experiment without compromising safety or consistency. Cao observed that this balance between access and control is becoming a central capability. Institutions that democratise AI effectively can accelerate learning and adoption across functions while sustaining the discipline required for responsible use. This shift places more responsibility on leadership teams to build governance structures that encourage participation rather than restrict it. These insights indicated that the next phase of AI adoption will be defined by how well organisations manage inclusion, oversight and alignment. New operating models favour agility and simplified architectures Institutions with simpler or more modern technology stacks described the clear advantages they have in integrating AI directly into core operations. Cao noted that these organisations can iterate faster, adopt end-to-end automation more naturally and embed intelligent decisioning into customer journeys without restructuring legacy systems. Several participants shared examples from emerging markets where agility, diverse data sources and modular architecture enable faster delivery of AI-enabled services. Cao remarked that such models often provide clearer demonstrations of what AI-first operating environments may eventually look like. These discussions underscored that future competitiveness will rely on an institution’s ability to adopt clean, adaptable architectures that support continuous learning and incremental improvement. Cao noted that this agility is not merely technical. It requires cultural alignment, clearer roles and accountability and a willingness to reorganise around new workflows. The contrast between legacy-heavy and digital-first organisations reinforced the importance of long-term architectural decisions. Collective dialogue shapes clearer direction for AI adoption Cao observed that the progression of the evening naturally supported deeper reflection. The departure from Sentosa Cove allowed for initial exchanges, while the pause between Lazarus Island and St John’s Island created an extended period for focused discussion. The route toward Pulau Bukom later in the evening reinforced the sense of scale and complexity in the challenges institutions face as they adopt AI. He reflected that the discussion demonstrated a growing willingness among leaders to confront the organisational implications of AI, not just the technical ones. Participants shared candid views on capability gaps, cultural barriers and the frameworks required to ensure responsible use. Cao noted that the dialogue highlighted how AI is prompting institutions to reconsider long-standing assumptions about control, efficiency and the distribution of decision-making authority. These reflections pointed to a more realistic and grounded understanding of what enterprise AI adoption entails. The conversations also showed that institutions increasingly see AI as a long-term capability rather than a sequence of discrete initiatives. Conclusion: clarity, governance and adaptability will shape the AI-ready institution The Singapore AI Sunset Cruise highlighted how institutions are preparing for a future where AI becomes an integral part of financial services. Cao’s observations underscored the need for clear enterprise direction, robust governance and the adaptability required to reorganise around intelligent systems. The dialogue showed that the institutions progressing most effectively are those treating AI as a structural change rather than a technical add-on. They are building the conditions for sustainable adoption by aligning culture, architecture and governance. The evening helped surface the questions that leaders must address as they define what an AI-ready institution should look like in the decade ahead.
