Find Insight addresses a core challenge in banking personalisation: instead of coarse segments by age and income, we use transactional and textual data to build behavioural profiles and to generate downstream product recommendations (insurance, lending, investments).
The core of the solution is an LLM-based semantic layer (“tags”) that converts heterogeneous events into personal tags and intents, thereby shifting from classical segmentation to hyper-segmentation at the individual level; on top of our domain ontology and product graph, a hybrid matching algorithm produces contextual, real-time recommendations.
The technological novelty is a proprietary, patentable approach to LLM tagging
The business impact includes a 40% uplift in cross/upsell and ARPU/LTV through higher relevance, up to a 30% reduction in CAC, and faster time-to-market for new offers—ultimately improving both the effectiveness of the product ecosystem and the customer experience.
