By Shavinya Abeywickrema and Catherine Yong
The GFTN Forum in Tokyo from 24-27 February, convened regulators, policy makers, startups, investors, and financial institutions, playing its role as the centrepiece event of Japan Fintech Week, in 2026. Organised by the not-for-profit Global Finance & Technology Network (GFTN), top thematic highlights at the four-day event revolved around financial corridors and capital flows and digital assets and tokenisation.
However, it was the agentic AI, governance and future tech stack (AI, quantum, infra) themes running across most sessions that stood out for Enterprise IT News.
On the final day, for example, a panel of top-tier investors from BEENEXT, Accion Venture Lab, Axion Ventures, and leadership from MUFG Group, discussed the evolving “World Order” of capital, the integration of AI and Blockchain into traditional banking, and Japan’s emergence as a strategic fundraising hub.
But AI and its adoption could not be overlooked and panellists observed that companies generally move through four stages of adoption – Productivity (headcount efficiency), Revenue Enhancement, Proprietary Data Monetisation, and full Transformation.
Enterprises are also undergoing AI Fatigue. AI-first pitches will not cut it anymore; enterprises now demand specific outcomes and would prefer white-label integrations to ensure ownership of the AI infrastructure/deployment.
Other notable observations of the discussions include India’s Digital Public Infrastructure (DPI) being positioned as the primary framework for financial services across the Global South, and how for the first time, institutions are disassociating blockchain technology from speculative crypto, focusing instead on Asset Tokenisation and B2B Stablecoins.
Standards for confidence
Later, during a FSA-SEC conversation, a big thread that emerged for EITN, was cross border alignment for digital market oversight.
Institutions need to be able to participate with more confidence and founders need to know what “good compliance” looks like. In essence, standards and enforcement have to be coordinated so as to reduce regulatory arbitrage and tackle fraud, with particular attention to areas that need common rules fast (stablecoins and tokenised real-world assets).
Kore.ai’s workshop titled “Enterprise AI without the chaos” by Sreeni Unnamatla, was a practical guide to getting enterprise AI out of the pilot stage and into production, by breaking rollout into clear phases: foundation setup, pilot design, expansion strategy, and scale-up governance – using examples from banks and insurers.
The core message was that AI programs stall for predictable reasons that aren’t due to the algorithm model and due to factors like fragmented data and systems, weak governance and trust, AI treated as an IT project instead of business transformation, unclear linkage to business outcomes, workflows not redesigned, too many pilots with no scaling discipline, and culture/talent/incentives barriers.
The recommended approach was equally pragmatic: build shared data foundations and MLOps, put responsible AI and lifecycle monitoring in place, secure executive sponsorship aligned to strategic outcomes, measure decision to outcome to value, redesign decision processes to embed AI, strengthen AI literacy and cross-functional teams, and treat AI as an enterprise capability that is repeatable, governed, and owned.
Proof, Not Promises: Test-and-Learn regulations for AI, payments and programmable money
The GFTN Insights roundtable was a closed-door session that was framed around a simple tension: AI is reshaping financial services faster than traditional rulemaking can keep up, not just because of new models, but because of new ways of deploying and combining existing techniques across products, channels, and infrastructure.
The discussion centred on “test-and-learn” approaches – sandboxes, pilots, evaluation methods, and shared experiments – as a way to generate evidence that builds trust without choking innovation. A specific case study used to ground the conversation was “agentic payments”: where AI is starting to influence payment workflows, raising practical questions about what enables safe scaling and which concerns matter most.
From there, the roundtable widened out across levels – from operational implementation to governance and policy – drawing parallels to other parts of finance, intersections with programmable money (DLT/tokenised payments), and emerging security horizons.
The intention was explicitly cross-jurisdictional: to surface lessons that travel across markets and sectors, in a conversational and forward-looking format.
