[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-86305-en":3,"doc-seo-86305-105":30,"detail-sidebar-cat-0-en-105":92},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":21,"is_downloadable":21,"audit_status":21,"page_count":22,"language":23,"language_code":24,"site_id":25,"html_lang":24,"table_of_contents":26,"faqs":27,"seo_title":13,"seo_description":14,"update_tm":28,"read_time":29},86305,13056703020460,"Valentina","https://ap-avatar.wpscdn.com/avatar/be000253dac470eee5d?_k=1778207105932848923",8,"Research & Report","Explainable Agentic System for Detection of Conversational Scams with Summary-Based Memory","Conversational scams increasingly threaten users by unfolding over weeks or months to build trust and extract money or sensitive information. Existing detectors mainly analyze isolated messages and struggle against this longer, evolving structure. This work extends single-message phishing detection into an explainable agentic system for conversation-level scam identification. It introduces ConScamBench-278 covering eight scam types and reports strong recall and accuracy, validated via user studies and usability testing.","An Explainable Agentic System for Detection of Conversational Scams with Summary-Based Memory  \narXiv :2607 . 11707v1 [ cs .MA] 13 Jul 2026  \nAhmed Omar Salim Adnan University of New Haven  \nM.S. in Data Science [aos. adnan01@cunh. newhaven. edu](aos. adnan01@cunh. newhaven. edu)  \nYogananda Manjunath University of New Haven  \nM.S. in Data Science [ymanj2@unh. newhaven. edu](ymanj2@unh. newhaven. edu)  \nShivanjali Khare  \nUniversity of New Haven  \nDepartment of Electrical &  \nComputer Engineering and Computer Science  \n[skhare@newhaven. edu](skhare@newhaven. edu)  \nJuly 14, 2026  \nAbstract  \nFollowing the rapid progress of generative Artificial Intelligence, there is a growing threat posed by conversational scams. These scams often span over multiple weeks or months, gradually build trust and request for money or sensitive information. Existing scam-detection systems mainly focus on isolated messages, which renders them inadequate against this evolving threat. This paper extends single-message phishing detection and presents an explainable agentic system for detecting sophisticated conversational scams. It also introduces ConScamBench-278, an initial public multi-category benchmark for conversational scam detection spanning eight scam types, released to support reproducible evaluation and future expansion. On isolated messages the single-message detector attains 100% phishing recall, while the conversation-level detector identifies all conversational scams in the public LoveFraud02 corpus (83/83) and reaches 97.8% accuracy (95% CI [95.4, 99.0]) on ConScamBench-278 . Two user studies (N = 100 and N = 45) further motivate the system: participants report frequently experiencing uncertainty when judging suspicious conversations. In an uncontrolled pre/post comparison, users self-reported trust, self-confidence, and perceived need for AI-based scam detection all increased (p \u003C 0.001 , Wilcoxon signed-rank) . The system also receives a System Usability Scale score of 74.7 (95% CI [72.5, 76.9]), above the established usability benchmark.  \n1 Introduction  \nThe emergence of cyberscams is not sudden; rather, it evolved from traditional fraud schemes. With time, the digital aspect of scams proliferated in a variety of ways and became more convincing, as they rely on the same historic principles of psychological manipulation [1] . One of the most convincing cyberscams, phishing, refers to the fraudulent practice of sending messages, often in the false disguise of reputable companies. Here, the intention is to convince the victim to reveal sensitive personal information required to get access to their finances [2] . In 2024, phishing or spoofing was the most reported cybercrime to the FBI, totaling 193,407 complaints [3] .  \nIn the modern era of computing and Artificial Intelligence (AI), these scams have evolved from simple email deception into multi-modal, AI-augmented attack ecosystems [4] . Conversational  \nscams are considered significantly more dangerous, as they can cause severe psychological trauma beyond financial harm [5] . Hazell et al. led an extensive study on modern techniques of AI-based phishing attacks and found that AI-generated emails were often more persuasive than human-crafted ones. According to [6], AI-generated phishing is the biggest enterprise threat of 2026, outpacing ransomware and insider risk. This concern is becoming more pressing as generative AI lowers the cost of producing fluent, personalized, and sustained scam conversations, allowing attackers to automate social-engineering messages that appear more natural than traditional template-based phishing. Given this progression, there is an immediate demand for equally strong countermeasures. Existing AI-based phishing detectors typically focus on detecting phishing in single messages;  \nwe refer to such a tool as a Single-Message Detector (SMD) . One such detection tool, SmishX, is a Large Language Model (LLM) based agentic phishing detector that takes a sin","cbCaikBY6clg7YYG","https://ap.wps.com/l/cbCaikBY6clg7YYG","pdf",800586,5,1,34,"English","en",105,"# Abstract\n# Introduction","[{\"question\":\"Why are conversational scams harder to detect than single-message phishing?\",\"answer\":\"Conversational scams evolve over weeks or months, gradually building trust and eliciting money or sensitive information. Many existing systems classify messages in isolation, missing delayed conversational signals.\"},{\"question\":\"What is ConScamBench-278 and what does it support?\",\"answer\":\"ConScamBench-278 is an initial public multi-category benchmark for conversational scam detection. It spans eight scam types to enable reproducible evaluation and future expansion.\"},{\"question\":\"How does the proposed system handle long conversations efficiently?\",\"answer\":\"It uses summary-based memory to maintain a compact representation of evolving conversations. This avoids repeatedly passing the full conversation history into the detector.\"}]",1784210348,86,{"code":4,"msg":31,"data":32},"ok",{"site_id":25,"language":24,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":87,"head_meta":89,"extra_data":91,"updated_unix":28},"explainable-agentic-system-for-detection-of-conversational-scams-with-summary-based-memory","",{"@graph":36,"@context":86},[37,54,69],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/explainable-agentic-system-for-detection-of-conversational-scams-with-summary-based-memory/86305/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":24,"description":14,"dateModified":62,"datePublished":63,"encodingFormat":61,"isAccessibleForFree":64,"interactionStatistic":65},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-07-25","2026-07-16",true,{"@type":66,"interactionType":67,"userInteractionCount":20},"InteractionCounter",{"@type":68},"ViewAction",{"@type":70,"mainEntity":71},"FAQPage",[72,78,82],{"name":73,"@type":74,"acceptedAnswer":75},"Why are conversational scams harder to detect than single-message phishing?","Question",{"text":76,"@type":77},"Conversational scams evolve over weeks or months, gradually building trust and eliciting money or sensitive information. Many existing systems classify messages in isolation, missing delayed conversational signals.","Answer",{"name":79,"@type":74,"acceptedAnswer":80},"What is ConScamBench-278 and what does it support?",{"text":81,"@type":77},"ConScamBench-278 is an initial public multi-category benchmark for conversational scam detection. It spans eight scam types to enable reproducible evaluation and future expansion.",{"name":83,"@type":74,"acceptedAnswer":84},"How does the proposed system handle long conversations efficiently?",{"text":85,"@type":77},"It uses summary-based memory to maintain a compact representation of evolving conversations. This avoids repeatedly passing the full conversation history into the detector.","https://schema.org",{"og:url":52,"og:type":88,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":90,"canonical":52},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":93},[94,98,102,106,110,115,120,123,128,131,135],{"id":21,"doc_module":4,"doc_module_name":46,"category_name":95,"show_sort_weight":96,"slug":97},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":99,"show_sort_weight":100,"slug":101},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":103,"show_sort_weight":104,"slug":105},"Exam",70,"exam",{"id":20,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":20,"slug":138},19,"General","general"]