Search papers, labs, and topics across Lattice.
FinTRACE, a retrieval-first architecture, is introduced to address the challenges of using general-purpose LLMs with time-distributed financial transaction data. It converts raw transactions into reusable feature representations, applies rule-based detectors, and stores signals in a behavioral knowledge base. FinTRACE significantly improves low-supervision transaction analytics, doubling zero-shot MCC on churn prediction and enabling state-of-the-art LLM performance through instruction tuning on retrieved behavioral patterns.
LLMs can now achieve state-of-the-art results on financial transaction analytics by grounding them with a retrieval-augmented knowledge base of behavioral patterns derived from raw transaction data.
Nowadays, success of financial organizations heavily depends on their ability to process digital traces generated by their clients, e.g., transaction histories, gathered from various sources to improve user modeling pipelines. As general-purpose LLMs struggle with time-distributed tabular data, production stacks still depend on specialized tabular and sequence models with limited transferability and need for labeled data. To address this, we introduce FinTRACE, a retrieval-first architecture that converts raw transactions into reusable feature representations, applies rule-based detectors, and stores the resulting signals in a behavioral knowledge base with graded associations to the objectives of downstream tasks. Across public and industrial benchmarks, FinTRACE substantially improves low-supervision transaction analytics, doubling zero-shot MCC on churn prediction performance from 0.19 to 0.38 and improving 16-shot MCC from 0.25 to 0.40. We further use FinTRACE to ground LLMs via instruction tuning on retrieved behavioral patterns, achieving state-of-the-art LLM results on transaction analytics problems.
