arXiv Computation and Language

CARRE: Counterfactual Action Retrieval and Reason Evaluation for Explainable Churn Prescription

arXiv AI
Aug 28

Explainable Artificial Intelligence for Customer Churn Prediction in Telecommunications: A Framework for CRM Integration

The paper presents a framework for integrating explainable AI into customer churn prediction for telecommunications. It benchmarks four classifiers—Logistic Regression, Random Forest, XGBoost, and LightGBM—on the IBM Telco Customer Churn dataset, finding comparable performance with Logistic Regression achieving the highest AUC-ROC and LightGBM the highest accuracy. Explanations are provided via SHAP and LIME at both global and instance levels, and a four‑layer CRM integration architecture is proposed to translate risk scores and attribution vectors into actionable retention strategies, projecting a 3.3–5.3 percentage point reduction in churn and $199K–$319K savings per campaign cycle.

By Sandeep Gaddamwar
arXiv Machine Learning
Oct 3

The Hidden Costs of 99% Accuracy: A Trustworthiness Audit of the Telco Customer Churn Benchmark

The paper audits the IBM Telco Customer Churn benchmark, revealing that common practices inflate performance metrics. It shows that pre‑split SMOTE boosts churn‑class F1 by 13.1 points, that isotonic regression is the best calibration method while temperature scaling fails on tree ensembles, and that the cost‑optimal decision threshold is 5–10 times lower than the F1‑optimal one, saving about $77,000 per 1,000 customers. The authors also test generalisation on Iranian Telecom and Bank churn datasets, and propose a four‑component reporting checklist with reproducible code.

By Soumyadeep Roy
arXiv Computation and Language
Aug 27

FinRiskAtlas: Decision-Aligned Evaluation of Large Language Models for Financial Risk Review

FinRiskAtlas is a Chinese-language benchmark designed to evaluate large language models (LLMs) for financial risk review by focusing on decision‑aligned tasks rather than generic financial knowledge. It contains 9,742 instances across 53 task families, including 42 domain‑knowledge families and 11 downstream review operations defined by explicit evaluation contracts. The extended FinRisk‑Ask framework replays 680 pre‑action states from 104 professional trajectories, withholding future evidence during inference to assess evidence‑state control and request targeting. Results across 33 model configurations show that operation‑level evaluation yields distinct rankings and that knowledge‑based shortlisting can incur significant regret, while frequent use of the Ask branch does not necessarily improve evidence acquisition, highlighting gaps in broad financial capability scores.

By Suyang Zhong, Jingzhe Zhu, Qi Xu, Liyao Sun, Yin Wang, Qingqing Sun, Shuai Chen, Tianyi Zhang
arXiv Machine Learning
Sep 25

From Prediction to Explainable Provider Behavior Profiles for Fraud, Waste, and Abuse Review

The paper proposes a shift from predictive modeling to descriptive provider behavior profiles for fraud, waste, and abuse (FWA) review. By decomposing billed revenue into provider scale and procedure composition, the authors construct lineage‑aware profiles that capture scale history, code lineage, and clinical family shares. In a large Medicare audit, these simple, interpretable descriptions outperform complex forecasts and improve recall for high‑cost rare events, while an optional semantic factorization adds context without inferring intent.

By Yubin Park, Evan Brociner
arXiv AI
Sep 25

Persuaded, Not Informed: Incentive-Misaligned Witnesses Defeat In-Context Grounding

The paper reports that language‑model agents used for customer‑relationship management can be misled by optimistic assertions from sales representatives in CRM records, leading to incorrect deal approvals. In a benchmark of 100 lead‑qualification tasks, models incorrectly cleared 29 of 31 deals where the representative’s claims contradicted company policies, with misalignment rates ranging from 87% to 97% across seven models. The authors propose diagnostic methods—including bucket analysis, same‑information controls, and compute‑step controls—to distinguish persuasion from information gaps and to quantify the impact of incentive‑misaligned witnesses.

By Rahul Balakavi
arXiv Computation and Language
Nov 5

CARTBench: Rubric-Based Evaluation of Shopping Advice for Complex and Ambiguous Requests with Timelines

CARTBench is a new benchmark for evaluating large language model shopping assistants on complex, ambiguous requests that involve timelines. It compiles 1,603 timestamped dialogues from 17 retail domains and pairs them with product records, then uses case‑specific rubrics based on Ideal Response Principles to score responses across Core, Excellence, and Safety criteria. In a full‑suite evaluation of ten commercial APIs and several open‑weight models, only three models scored above 90% on Core and none surpassed 45% on Excellence, revealing a gap between required behavior and high‑quality purchase advice.

By Yuqin Dai, Zhanwei Zhang, Honghui Yang, Zhouheng Yao, Yueqi Zhang, Jingyu He, Chaoyang Shi, Fanding Wei, Boyuan Pan, Kai Xu, Yao Hu, Yongliang Shen
arXiv AI
Aug 11

FinTrace: Holistic Trajectory-Level Evaluation of LLM Tool Calling for Long-Horizon Financial Tasks

arXiv:2604. 10015v3 Announce Type: replace Abstract: Recent studies demonstrate that tool-calling capability enables large language models (LLMs) to interact with external environments for long-horizon financial tasks.

By Yupeng Cao, Haohang Li, Weijin Liu, Wenbo Cao, Anke Xu, Lingfei Qian, Xueqing Peng, Minxue Tang, Zhiyuan Yao, Jimin Huang, K. P. Subbalakshmi, Zining Zhu, Jordan W. Suchow, Yangyang Yu