Financial Services
Duna says deterministic rules are key to bank-grade AI
Duna is promoting rule-based orchestration of AI agents for KYC, screening, monitoring, and due diligence, addressing reliability concerns in regulated finance.
An analysis of AI adoption in financial services highlighted the use of deterministic rules to make agentic systems more reliable in regulated banking workflows. Duna’s policy engine converts a bank’s policies into code and coordinates specialized AI agents across onboarding, due diligence, screening, monitoring, and ongoing know-your-customer processes. These are high-control activities where inconsistent outputs can create compliance and operational risk. The report said 70% of financial-services firms and regulators rank model hallucinations and unreliable outputs among their top AI risks. Duna’s approach is designed to place explicit policy logic around AI agents so that automated work remains aligned with institutional rules. The application is relevant to MENA financial institutions because KYC, sanctions screening, and transaction monitoring are common requirements across the region’s banking systems. The source describes the product and the broader risk findings but does not identify a specific MENA customer or report quantified production results. The development reflects a wider industry shift from general-purpose generative AI experimentation toward controlled automation for compliance and other regulated workflows.
Original source: FinTech Global