Financial institutions in Southeast Asia are rapidly embedding artificial intelligence into live credit risk and fraud detection systems, but the real challenge is no longer technological capability: governance.
As the region moves from experimentation to full-scale deployment of AI in financial decision-making, banks and fintechs are increasingly confronting a more complex question: how to maintain control, accountability, and explainability when agentic AI begins to act in real time.
Against this backdrop, Singapore-headquartered risk intelligence provider TrustDecision has expanded its Agentic AI capabilities for banks and fintechs across Southeast Asia.
The upgraded platform introduces domain-specific AI agents designed to enable end-to-end, closed-loop risk decisioning — from financial fraud detection and investigation to strategy optimization.
The announcement was made during Money20/20 Asia in Bangkok this week, where the company positioned its vision as an “AI-native decisioning infrastructure” for financial services.
The platform is already used by institutions including Hong Leong Bank, Mandiri, Bank Jago, Toyota Auto Finance (TAF), Kredivo, Adakami, among many others.
From automation to agentic decisioning
TrustDecision’s latest expansion reflects a broader industry shift away from rule-based systems and even conventional AI models, toward what it calls agentic AI — systems capable of performing structured tasks across the entire risk lifecycle.
At the core of this approach are domain-specific AI agents, including:
- Investigation AI agents, which support fraud and money laundering case work through automated data retrieval, fund tracing, behavioural analysis, and report generation. The aim is to accelerate investigations while maintaining consistency and human-AI collaboration.
- Rule mining AI agents, which translate investigation outcomes and risk signals into deployable rules and model features, allowing institutions to move from reactive case handling to continuously improving risk strategies.
Together, these agents form a closed-loop system linking detection, investigation, and strategy optimization — designed to make risk decisioning more adaptive and scalable.
Governance becomes the bottleneck
Henry Li Nan, Managing Director for Singapore, Malaysia & Thailand at TrustDecision
However, as adoption accelerates, financial institutions are increasingly running into a shared constraint: governance.
While autonomous AI systems can improve efficiency and reduce operational costs, they also introduce new challenges around auditability, compliance, and accountability — particularly in regulated financial environments.
TrustDecision argues that the focus is shifting from what AI can do to how it is deployed and controlled in production environments.
According to the company, AI agents are best suited for high-volume, lower-risk processes such as transaction flagging and application pre-scoring, while higher-stakes financial decisions still require clear human oversight.
Henry Li Nan, Managing Director for Singapore, Malaysia & Thailand at TrustDecision, said banks are taking a more cautious, structured approach to adoption.
“Many incumbent banks are actively exploring AI today, and the pace of technological advancement has been faster than expected. But in regulated financial environments, the focus is not just autonomy, it’s on control,” he said. “While AI agents can take on more operational tasks, institutions are adopting a measured approach, with human-AI collaboration at the core. Higher-risk decisions still require human oversight, and there is a strong emphasis on model transparency and operating within clearly defined regulatory boundaries.”
Thailand’s virtual bank rollout raises urgency
Dr. Simon Liu, Chief Data and AI Officer at TrustDecision
Thailand is emerging as a key testing ground for this shift. With its first batch of virtual banks expected by mid-2026, financial institutions are under pressure to build digital-first systems that are both efficient and compliant from the outset.
This transition is creating stronger demand for AI-native infrastructure — but also heightening scrutiny around transparency, governance, and system accountability.
Across Southeast Asia, adoption remains uneven. Some markets are moving quickly into real-time decisioning and digital banking, while others are still developing foundational infrastructure.
However, TrustDecision notes a common shift in mindset: from whether AI should be used, to how it can be safely controlled at scale.
Dr. Simon Liu, Chief Data and AI Officer at TrustDecision, said the industry has largely moved past capability concerns.
“The capability questions are largely answered. AI can make faster, more consistent decisions than manual processes in almost every financial workflow we handle,” he said. “In regulated financial services, the more important question is how the system behaves once it is live. That means understanding where autonomous decision is appropriate, where controls need to sit, and how institutions detect when a system is operating outside the boundaries they have set.”
From experimentation to controlled deployment
TrustDecision’s perspective was also presented at Money20/20 Asia, where Dr. Liu joined a panel on agentic finance and AI as a financial actor, while Henry Li Nan participated in the Policy20 Roundtable, a closed-door session involving regulators and policy stakeholders across Asia-Pacific.
As financial institutions transition from pilot projects to production-scale deployment, TrustDecision believes the next phase of AI adoption will depend less on model performance — and more on whether institutions can build governance systems capable of managing AI safely in live financial environments.
