AI is no longer a pilot project in financial services. That was the clear message at the 2026 BAFT International Trade & Payment Conference, where a panel on AI in compliance and fraud detection explored what comes next for banks and payments providers.
According to Quantifind, moderated by BNY senior vice president of global payments and trade Ryan Lastra, the discussion brought together Quantifind VP of strategic client partnerships Teresa Buechner and BNY senior vice president of domestic payments Sumner Francisco.
What stood out was not a debate over whether AI belongs in financial services. That argument has already been settled. Instead, the focus shifted to governance, explainability and collaboration in a world where financial crime operates across networks, not institutions.
AI is already embedded in payments monitoring, fraud detection and compliance workflows. The pressing challenge now is whether institutions can manage and scale it responsibly across an ecosystem where risk flows between banks, platforms and payment rails.
For years, conversations centred on potential. What could AI achieve? When would regulators allow it? Where might it fit? Today, AI systems are actively monitoring transactions in real time, flagging anomalies and helping teams tackle enormous alert volumes.
Buechner highlighted examples where AI rescored legacy alerts and cleared nearly 70% of false positives, transforming hundreds of thousands of cases into a more focused, high-risk subset. The operational gains are clear.
Yet the panel stressed that internal optimisation is no longer sufficient. Risk does not sit neatly within a single account or institution. It exists within networks — shaped by counterparties, transaction pathways and behavioural patterns that stretch across the financial system.
A customer may appear low risk when viewed in isolation, but be deeply connected to high-risk activity elsewhere. Fraudsters understand this dynamic and exploit the visibility gaps between institutions.
AI can act as a powerful force multiplier, but only when it has access to the right data. Machine learning models surface patterns at scale. Generative AI can accelerate investigations and summarise complex case files. However, humans still define risk appetite and make final decisions.
The augmentation model breaks down when data is fragmented. Both Buechner and Francisco underscored that AI’s effectiveness depends on the breadth and quality of signals available. When institutions operate in silos, even sophisticated models are constrained. When insights can be shared securely and within regulatory guardrails, AI’s impact increases significantly.
The discussion also highlighted payment rails as a potential collaboration layer. Infrastructure operated by central bodies effectively sits at the heart of transaction flows. These shared rails represent natural aggregation points where cross-institution risk patterns emerge. Used effectively, they could support earlier detection of coordinated fraud and deliver more consistent responses across the industry. Much of the infrastructure already exists; the remaining challenge lies in aligning governance frameworks and incentives.
The panel also challenged the perception that regulators are barriers to AI adoption. The concern is not the use of AI itself, but opacity. Supervisors expect explainability, traceability and robust model governance. Institutions that understand and document their decision pathways are better positioned to innovate and to participate in responsible data-sharing initiatives.
Fraudsters already operate collectively, using AI to probe systems and adapt rapidly. Defending against that threat requires shared intelligence and faster feedback loops. The next two years, the panel suggested, will be defined by convergence — fraud, AML, cyber and payments risk merging into one interconnected challenge.
AI may be embedded everywhere, but governance and collaboration will determine whether it truly strengthens the financial ecosystem.
Copyright © 2026 RegTech Analyst
Copyright © 2018 RegTech Analyst
