[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-82558-en":3,"doc-seo-82558-105":30,"detail-sidebar-cat-0-en-105":91},{"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},82558,962075006959,"Anda","https://ap-avatar.wpscdn.com/avatar/e0002397efbe92a78e?_k=1776741047341049297",8,"Research & Report","Exploring the Semantic Gap in Agentic Data Systems","Large language models increasingly generate queries, call tools, and build analytical workflows, yet the semantic information needed to operationalize analytical concepts often is not represented in database schemas or stored values. This cross-domain formative study analyzes 236 analytical intents across finance, human resources, and public safety and finds 153 recurring operationalization failures. Five failure classes emerge: comparative grounding, process reasoning, quantitative reasoning, role confusion, and policy grounding, indicating a semantic gap affecting analytical admissibility.","Exploring the Semantic Gap in Agentic Data Systems: A Formative Study of Operationalization Failures in Analytical Workflows  \nJalal Mahmud  \nMegagon Labs California, USA [jalal@megagon.ai](jalal@megagon.ai)  \nEser Kandogan  \nMegagon Labs California, USA [eser@megagon.ai](eser@megagon.ai)  \narXiv :2607 .00828v 1 [ cs .DB] 1 Jul 2026  \nABSTRACT  \nLarge language models (LLMs) are increasingly used to generate queries, invoke tools, and construct analytical workflows. Although recent advances have substantially improved workflow generation and execution, the semantic information required to operationalize analytical concepts often lies beyond what is explicitly represented in database schemas and data values. We present a cross-domain formative study of operationalization failures in agent-generated analytical workflows. Across 236 analytical intents spanning finance, human resources, and public safety domains, we identify 153 recurring failures despite successful workflow generation and execution. Our analysis reveals five recurring classes of failures: comparative grounding, process reasoning, quantitative reasoning, role confusion, and policy grounding. These findings suggest a semantic gap between user-level analytical concepts and the information available to workflow-generation systems. More broadly, they raise questions about the admissibility of analytical operations and suggest that future agentic data systems may require richer semantic representations to bridge the gap between analytical intent and executable computation.  \n1 INTRODUCTION  \nLarge language models (LLMs) are rapidly becoming a core primitive of modern data systems. Modern AI-native systems increasingly rely on LLMs to generate queries, retrieve evidence, invoke tools, and construct analytical workflows. Examples include NL2SQL systems, retrieval-augmented systems, enterprise analytical assistants, and agentic planners that combine reasoning with database and retrieval operations [7, 8, 14, 17, 18] . Recent advances have substantially improved intent interpretation, workflow generation, and answer correctness. Nevertheless, analytical failures continue to occur even in workflows that execute successfully.  \nConsider an analytical assistant tasked with answering the request: ‘Find unusually risky loans with sustained delinquency.” A generated workflow may operationalize ‘unusually risky” using a fixed threshold and ‘sustained delinquency” using a row-level predicate. However, unusual risk is inherently comparative and typically requires reasoning relative to a reference population, while sustained delinquency describes a temporal process that cannot be inferred from a single observation. Similarly, when asked to‘identify intersections with the highest fatality rate,” the generated workflow may rank intersections using total fatalities rather than normalized fatality rates. Such a workflow conflates counts with rates, favoring high-volume intersections even when their fatality rate is lower. In both cases, the generated workflow is syntactically valid and executable, yet the selected operations do not faithfully capture the analytical concept expressed in the user’s intent.  \nThese examples suggest a broader semantic gap in agentic data interaction. Existing data systems have traditionally focused on bridging gaps in intent interpretation, schema understanding, and query generation. Our findings suggest an additional semantic frontier concerned with the operationalization of analytical concepts. Concepts such as unusual, persistent, high-risk, and rate must ultimately be translated into executable computations, yet the information required to perform this translation is often not explicitly represented in database schemas or data values.  \nAs shown in Figure 1, human analysts routinely draw on statistical context, process knowledge, metric definitions, analytical roles, and organizational policies when interpreting such concepts, whereas agents often have a","cbCaifdDVQRBh3xZ","https://ap.wps.com/l/cbCaifdDVQRBh3xZ","pdf",1899221,3,1,5,"English","en",105,"# Abstract\n# Introduction\n# Related Work","[{\"question\":\"What problem does the study investigate in agentic analytical workflows?\",\"answer\":\"It examines operationalization failures where agent-generated workflows execute successfully but fail to faithfully capture the analytical concepts expressed in user intent.\"},{\"question\":\"How was the study conducted and what scope did it cover?\",\"answer\":\"The study analyzes 236 analytical intents across finance, human resources, and public safety, identifying 153 recurring failures even when workflow generation and execution succeed.\"},{\"question\":\"What are the five recurring classes of operationalization failures?\",\"answer\":\"The paper reports comparative grounding, process reasoning, quantitative reasoning, role confusion, and policy grounding.\"}]",1784181522,13,{"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":86,"head_meta":88,"extra_data":90,"updated_unix":28},"exploring-the-semantic-gap-in-agentic-data-systems","",{"@graph":36,"@context":85},[37,53,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,50],{"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":20},"https://docshare.wps.com/document/research-report/",{"item":51,"name":13,"@type":43,"position":52},"https://docshare.wps.com/document/exploring-the-semantic-gap-in-agentic-data-systems/82558/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":24,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":41,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-23","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does the study investigate in agentic analytical workflows?","Question",{"text":75,"@type":76},"It examines operationalization failures where agent-generated workflows execute successfully but fail to faithfully capture the analytical concepts expressed in user intent.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How was the study conducted and what scope did it cover?",{"text":80,"@type":76},"The study analyzes 236 analytical intents across finance, human resources, and public safety, identifying 153 recurring failures even when workflow generation and execution succeed.",{"name":82,"@type":73,"acceptedAnswer":83},"What are the five recurring classes of operationalization failures?",{"text":84,"@type":76},"The paper reports comparative grounding, process reasoning, quantitative reasoning, role confusion, and policy grounding.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":92},[93,97,101,105,109,114,119,122,127,130,134],{"id":21,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":22,"doc_module":4,"doc_module_name":46,"category_name":106,"show_sort_weight":107,"slug":108},"Comic",60,"comic",{"id":110,"doc_module":4,"doc_module_name":46,"category_name":111,"show_sort_weight":112,"slug":113},6,"Technology",50,"technology",{"id":115,"doc_module":4,"doc_module_name":46,"category_name":116,"show_sort_weight":117,"slug":118},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":120,"slug":121},30,"research-report",{"id":123,"doc_module":4,"doc_module_name":46,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":46,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":22,"slug":137},19,"General","general"]