# SAGE: An LLM-driven Self Reflective Agentic Framework for Fraud Detection

arXiv: https://arxiv.org/abs/2606.08146 · HTML: https://mlsanity.com/paper/2606.08146/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Agent threats, guardrails and governance, one of 16 directions in LLM agents and tool use.
- Builds on 18 of its 45 references that are in the map.
- First posted on arXiv 6 Jun 2026.

## Placed in

- LLM agents and tool use / [Agent threats, guardrails and governance](https://mlsanity.com/agents/agent-threats-guardrails-and-governance/)

## Builds on (most cited first, 18 of 18)

- [ReAct: Synergizing Reasoning and Acting in Language Models](https://mlsanity.com/paper/2210.03629/) (2022)
- [XGBoost: A Scalable Tree Boosting System](https://mlsanity.com/paper/1603.02754/) (2016)
- [Reflexion: Language Agents with Verbal Reinforcement Learning](https://mlsanity.com/paper/2303.11366/) (2023)
- [Automatic Prompt Optimization with "Gradient Descent" and Beam Search](https://mlsanity.com/paper/2305.03495/) (2023)
- [Tabular Data: Deep Learning is Not All You Need](https://mlsanity.com/paper/2106.03253/) (2021)
- [AutoGluon-Tabular: Robust and Accurate AutoML for Structured Data](https://mlsanity.com/paper/2003.06505/) (2020)
- [Data Interpreter: An LLM Agent For Data Science](https://mlsanity.com/paper/2402.18679/) (2024)
- [Anti-Money Laundering in Bitcoin: Experimenting with Graph Convolutional Networks for Financial Forensics](https://mlsanity.com/paper/1908.02591/) (2019)
- [A Semi-supervised Graph Attentive Network for Financial Fraud Detection](https://mlsanity.com/paper/2003.01171/) (2020)
- [FLAML: A Fast and Lightweight AutoML Library](https://mlsanity.com/paper/1911.04706/) (2019)
- [Digital Deception: Generative Artificial Intelligence in Social Engineering and Phishing](https://mlsanity.com/paper/2310.13715/) (2023)
- [Semi-supervised Credit Card Fraud Detection via Attribute-Driven Graph Representation](https://mlsanity.com/paper/2412.18287/) (2024)
- [Graph Neural Networks for Financial Fraud Detection: A Review](https://mlsanity.com/paper/2411.05815/) (2024)
- [Credit card fraud detection using machine learning: A survey](https://mlsanity.com/paper/2010.06479/) (2020)
- [Fraud Dataset Benchmark and Applications](https://mlsanity.com/paper/2208.14417/) (2022)
- [VecAug: Unveiling Camouflaged Frauds with Cohort Augmentation for Enhanced Detection](https://mlsanity.com/paper/2408.00513/) (2024)
- [A Comprehensive Survey on Machine Learning Techniques and User Authentication Approaches for Credit Card Fraud Detection](https://mlsanity.com/paper/1912.02629/) (2019)
- [FactGuard: Event-Centric and Commonsense-Guided Fake News Detection](https://mlsanity.com/paper/2511.10281/) (2025)

## Built on by (most cited first, 1 of 1)

- [Toward Auditable Fraud Detection: Combining Graph Features, Model Explanations, and Agentic Case Investigation](https://mlsanity.com/paper/2607.19266/) (2026)

## Abstract (arXiv)

Fraud detection in payment, e-commerce, and telecommunications systems requires accuracy at the individual level, robustness under severe class imbalance, and ease of understanding for risk managers. Existing methods fall at least one of these requirements: automated machine learning systems search a fixed numerical space without semantic awareness of the dataset; graph neural network-based methods require pre-defined relational graphs and remain opaque at the individual-decision level; and the design of general-purpose large language model (LLM) agents does not consider the recall and precision constraints specific to real-world fraud detection. In this paper, we propose SAGE, the first end-to-end LLM-driven multi-agent framework for fraud detection. SAGE coordinates three dedicated agents that make decisions based on a six-layer Data Diagnostic Tree (DDT) and a Markov decision process guided by natural-language gradients, automatically optimizing the model under a fraud-specific reward. On five fraud datasets and five LLM backbones, SAGE wins 96.00% of method--dataset comparisons and improves F1 by an average of 40.86% over baselines. The code is available at https://github.com/yichenC1c/SAGE.