New Product Approval Risk Assessment

The Problem

Banks launch new products regularly (new credit product, new payment service, new investment vehicle, new customer segment). Regulatory requirement: comprehensive risk assessment before product launch. Current process: product team submits product specification; risk committee gathers: product design details, target customer segment, risk taxonomy applicability (credit, market, operational, reputational), regulatory requirements, pricing/profitability model, competitive landscape. Risk team manually maps product to risk domains, identifies regulatory requirements, generates preliminary risk assessment. Time to complete: 10-20 hours of risk team effort over 2-4 weeks. Process is often bottlenecked by risk team availability and product team's late risk documentation. Products sometimes launch with incomplete risk assessment, creating downstream regulatory risk.

Data Requirements

Data Sources:

• Product Specification Form (internal) , Product name, description, customer target, product structure, operational model, pricing, distribution strategy
• Risk Taxonomy Database (internal) , Risk domain definitions, risk classification matrix by product type, risk assessment criteria
• Regulatory Database (internal maintained) , Applicable regulations by jurisdiction and product type (MAS, EBA, OCC, CFPB, SEC, etc.), regulatory requirements, exemptions, safe harbors
• Product Library / Prior Risk Assessments (internal) , Historical product launches, prior risk assessments, product features, risk mitigations applied
• Capital Model Database (internal) , Capital requirements by risk domain (credit, market, operational risk), stress testing assumptions
• Risk Appetite Framework (internal) , Board-approved risk limits, risk tolerance by product type, risk return targets
• Compliance/Regulatory Change Database , Recent regulatory updates relevant to new product type, enforcement actions, guidance documents

Data Classification:

• Public: Published regulations (MAS, EBA, OCC, CFPB, SEC), recent regulatory announcements
• Internal Structured: Product specifications, risk taxonomy, regulatory requirement checklist, risk assessment templates, capital models, risk appetite limits
• Internal Unstructured: Prior risk assessment narratives, regulatory correspondence, product committee minutes
• External Third-Party: Regulatory guidance documents, published risk frameworks
• Sensitive/Regulated: Product specifications (confidential), risk assessments (can influence product decisions), regulatory submissions

Data Quality Requirements:

Product specification completeness: 95% (some details may be incomplete during early stages). Risk taxonomy consistency: 100% (taxonomy must be stable and well-documented). Regulatory database freshness: quarterly updates to reflect regulatory changes. Historical product data completeness: 80%+ (some legacy products may have incomplete records). Capital model accuracy: 100% reconciliation with regulatory capital calculations.

Integration Complexity: Medium , Requires access to product specification forms (may be in SharePoint, Jira, custom system). Risk taxonomy maintenance (may be manual or in PolicyTech system). Regulatory database integration (requires regular updates from regulatory agencies). Capital model integration (may require direct integration with risk modeling systems or APIs). Product lifecycle tracking (may require integration with project management tool). Output is primarily informational support for human risk committee review (lower automation complexity than decision-making use cases).

Regulatory Alignment

• BCBS 239 (Risk Governance): Agent supports risk assessment documentation required by Basel Committee
• Dodd-Frank (Systemically Important Financial Institutions): New product risk assessment required for SIFIs
• MAS Notice 608 (Interest Rate Risk Management in Banking Book): Agent assesses interest rate risk for deposit products
• GLBA (Product Governance): Agent supports risk-based product approval process

Comparable Implementations

• Citi: Product risk assessment automation; accelerates product committee review; 30% reduction in assessment time
• BNY Mellon: New product approval workflow integration; standardized risk taxonomy application