Galadrim enhanced an AI-driven advisory engine to construct equity and money-market portfolios for BNP Paribas and Hello bank! clients.
+15% efficiency across advisory operations
6 months to deliver measurable improvements
9/10 stakeholder satisfaction
BNP Paribas required a sharper, faster advisory algorithm to recommend suitable investment products (such as life insurance policies and Stocks & Shares ISAs) based on preferences captured via a client questionnaire. The priority was to boost speed, reliability, and suitability while preserving auditability.
We focused on targeted improvements to the algorithm and its decision pathway :
Data integrity and suitability: Validate risk appetite and investment amounts, enforce guardrails, and standardise inputs so recommendations consistently reflect client profiles.
Portfolio construction logic: Refine derived variables and asset-selection rules across equity and money-market universes to deliver more consistent, comparable portfolios.
Performance and observability: Profile the advisory flow, remove bottlenecks to meet throughput/latency targets, and strengthen monitoring for proactive issue detection.
Heterogeneous product set: Unified rules across life insurance, ISAs, and fund recommendations.
Consistency at scale: Delivered deterministic outcomes with full traceability for audits.
Live service constraints: Shipped improvements without disrupting client-facing journeys.
