Case Study: Belvo unlocks SMB lending and underwriting at scale with Ntropy

Belvo boosts transaction classification to 94.62% with Ntropy

Belvo, a major open banking provider in Europe, needed to help its customers in the SMB lending space make fast, standardized credit decisions. Their challenge was activating and enriching bank transaction data to enable profitable small business underwriting at scale, a process traditionally seen as operationally unjustifiable for banks.

Ntropy implemented its LLM-powered solution, using claude-3.5-sonnet and a caching infrastructure, to dynamically categorize transactions and calculate P&L from bank feeds. The results were a 94.62% accuracy in P&L labeling and a massive reduction in cost and latency. For Belvo's volume of 30 million transactions per day, Ntropy achieved a 155x cost reduction and a 163x latency reduction, unlocking a new capability for Belvo to expand its customer base into commercial banking.

Belvo