Data quality has emerged as the leading concern in AI Compliance in FinTech for financial institutions assessing artificial intelligence, AI-driven systems and machine learning algorithms for regulatory and financial-crime controls, according to a survey by Risk.net and Fenergo. The findings, reported by Asian Banking & Finance, are based on responses from 110 practitioners at banks and asset managers in Singapore, Malaysia and Australia.
The results place data governance at the centre of financial institutions’ AI planning. The report identified legacy systems, fragmented data and manual processes as barriers to implementation, alongside governance challenges and shortages of staff with the necessary technical skills. These issues can affect business verification, fraud detection and anti-money laundering controls, particularly where financial organizations rely on automated decision-making.
The finding is consistent with the direction of emerging AI requirements. The European Commission’s AI Act implementation guidance states that high-risk AI systems will require high-quality datasets, risk assessment and mitigation, activity logging, documentation, human oversight, robustness, cybersecurity and accuracy. The Act became applicable on 2 August 2026, while certain high-risk obligations will apply later under the current implementation timeline.
For organizations reviewing their data controls, Nemko Digital’s overview of ISO/IEC 5259-4 and data quality for machine learning describes a framework covering data acquisition, composition, preparation, labelling, evaluation and use. These controls are relevant to both model development and the evidence needed to assess AI performance over time, support greater accuracy and reduce potential risks.
Despite the concerns, AI adoption remains under consideration across the surveyed institutions. Generative AI was cited by 77% of respondents, followed by machine learning at 68%. Robotic process automation was cited by 47%, natural language processing by 46% and agentic AI by 44%. Only 2% said their organizations were considering no forms of AI.
Among respondents considering agentic AI, transaction monitoring was the leading intended use case at 66%. Fraud detection followed at 55%, sanctions screening at 46%, KYC maintenance at 42% and customer onboarding at 41%. These applications may depend on identity proofing (idp), AI-driven KYB solutions and reliable detection of suspicious activity and suspicious patterns.
Singapore’s Monetary Authority has also identified lifecycle controls as a supervisory priority. In its November 2025 proposed Guidelines on AI Risk Management, MAS said financial institutions should maintain accurate, up-to-date AI inventories and apply controls covering data management, fairness, transparency, human oversight, third-party risk, testing, monitoring and change management. The proposed guidance covers generative AI and newer developments such as AI agents, while highlighting the need to address AI bias and meet evolving regulatory changes.
For compliance, risk and technology teams, the survey highlights the need to establish where AI is used, what data supports each application and how results can be traced and challenged. Organizations should also monitor regulatory implementation, particularly requirements affecting high-risk systems and general-purpose AI, while assessing whether existing governance processes can accommodate autonomous or semi-autonomous agents. This includes reviewing cybersecurity controls against relevant information security guidance.
Compliance teams and compliance officers should also review AML processes, compliance protocols and compliance management practices to ensure they support robust regulatory adherence. This includes considering data privacy rules, financial crime regulations, AML regulations, AI regulatory requirements and automated regulatory reporting, while maintaining appropriate ethical standards and fairness.
Nemko Digital’s resources on AI laws for businesses and building trust in AI provide further context on the expanding compliance and governance landscape. The immediate issue identified by the survey, however, is operational: before financial firms can rely on AI for sensitive compliance decisions, they must be able to demonstrate that the underlying data is fit for purpose, controlled and continuously monitored. Doing so can improve operational efficiency, customer service and accountability, while giving organizations a strategic advantage as AI innovation continues.